metadata: run_metadata: 1,851 rows where experiment.library_selection = "cDNA", technology = "smartseq" and tissue

1,851 rows where experiment.library_selection = "cDNA", technology = "smartseq" and tissue_curation = "Embryo Imprecise"

descending

Link	rowid ▼	run.accession	experiment.accession	sample.accession	study.accession	bioproject	study.title	study.alias	study.type	study.abstract	study.PMIDs	sample.title	sample.alias	sample.attributes	GEOsample.title	GEOsample.dataprocessing	GEOsample.source	GEOsample.treatmentprotocol	GEOsample.extractprotocol	GEOsample.growthprotocol	GEOsample.characteristics	GEOsample.accession	experiment.title	experiment.alias	experiment.library_name	experiment.design_description	experiment.library_construction_protocol	experiment.attributes	experiment.library_strategy	experiment.library_source	experiment.library_selection	experiment.library_layout	experiment.platform	experiment.instrument_model	experiment.study_ref	run.attributes	run.filename	run.semantic_name	run.total_bases	run.total_spots	run.alias	run.read_lengths	run.base_counts	run.r1_length	run.r2_length	run.Acount	run.Ccount	run.Gcount	run.Tcount	run.Ncount	run.experiment	run.pool_member	submission.accession	submission.srasource	submission.bioprojectsource	seqdetective.n_mates	seqdetective.mapping_rate.mate1	seqdetective.mapping_rate.mate2	seqdetective.nofeature_rate.mate1	seqdetective.nofeature_rate.mate2	seqdetective.sparsity.mate1	seqdetective.sparsity.mate2	seqdetective.pos_strand_rate.mate1	seqdetective.pos_strand_rate.mate2	seqdetective.readlen.mate1	seqdetective.readlen.mate2	seqdetective.judgement.mate1	seqdetective.judgement.mate2	seqdetective.judgement.reason	platform_family	instrument_generation	read_bias	selection_class	prep_kit	sc_or_bulk	tech_class	technology	submission.bioprojectsource.country	earliest_date	devstage_curation	devstage_curation_coarse	tissue_curation	tissue_curation_coarse
29202	29202	SRR27329257	SRX23006269	SRS19970080	SRP479683	PRJNA1055904	BMP dependent patterning of ectoderm tissue material properties modulates lateral mesendoderm cell migration during early zebrafish gastrulation	GSE251904	Transcriptome Analysis	Cell migration is a fundamental process during embryonic development. Most studies in vivo have focussed on the migration of cells using the extracellular matrix ECM as their substrate for migration. In contrast much less is known about how cells migrate on other cells as found in early embryos when the ECM has not yet formed. Here we show that lateral mesendoderm LME cells in the early zebrafish gastrula use the ectoderm as their substrate for migration. We show that the lateral ectoderm is permissive for the animal pole directed migration of LME cells while the ectoderm at the animal pole halts it. These differences in the permissiveness depend on the lateral ectoderm being more cohesive than the animal ectoderm a property controlled by BMP signalling within the ectoderm. Collectively these findings identify ectoderm tissue cohesion as one critical factor regulating LME migration during zebrafish gastrulation. Overall design: To investigate the differences in gene expression between animal and lateral ectoderm causing the difference in material properties between the two ectodermal tissues.	pubmed:40057955	Zebrafish Lateral Ectoderm 7.5 hpf 4	GSM7989708	source name:Lateral Ectoderm\|tissue:Lateral Ectoderm\|genotype:WT TgSebox::eGFP embryos\|treatment:1 cell stage injection of PAmCherry1 mRNA Photoactivation FACS\|geo loc name:missing\|collection date:missing	Zebrafish Lateral Ectoderm 7.5 hpf 4	FastQC quality control Trimmomatic adapter & quality trimming; parameters: 2:30:10 LEADING:3 TRAILING:3 SLIDINGWINDOW:4:15 MINLEN:19 FastQC controlling trimming results salmon Alignment; used flags: seqBias qcBias R DESeq2 differential gene expression analysis; DESEq2 apeglm tximport genefilter GenomicRanges R EnhancedVolcano & biomaRt visualization Assembly: Danio rerio GRCz11 Supplementary files format and content: tab delimited text file with gene name gene length effective length tpm and number of reads for each sample	Lateral Ectoderm	Embryo were injected with 125 pg of PAmCherry1 mRNA at xxx cell stage. At 6 hpf embryos were mounted for up right imaging and a 444 x 220 x 58.5 µm volume was photoactivated using a 405 nm laser either at the animal pole or on the lateral side. The photoactivation process was composed of 10 cycles of 28 sec per embryo in a time window of approximately 90 min. post photoactivation embryos were transferred in Ca2+ free Ringer solution and dissociated by gently pipetting using a P 1000 pipette to obtain a single cell suspension. 5000 PAmCherry1+ cells were sorted in 150 µl of RTL plus lysis buffer RNeasy Plus Micro Kit QIAGEN with 1% of 2 mercaptoethanol.	Cells were subjected to RNA extraction using RNeasy Plus Micro Kit QIAGEN according to the manufacturer instructions Complete cDNA synthesis and library preparation were performed using SMART Seq v3 protocol with Nextera UDI adapters. Libraries were then quantified by qPCR KAPA Biosysytems.	Embryos were kept at 28.5 C until dissociation	tissue:Lateral Ectoderm\|genotype:WT TgSebox::eGFP embryos\|treatment:1 cell stage injection of PAmCherry1 mRNA Photoactivation FACS	GSM7989708	GSM7989708: Zebrafish Lateral Ectoderm 7.5 hpf 4; Danio rerio; RNA Seq	GSM7989708 r1	GSM7989708	1	Cells were subjected to RNA extraction using RNeasy Plus Micro Kit QIAGEN according to the manufacturer instructions Complete cDNA synthesis and library preparation were performed using SMART Seq v3 protocol with Nextera UDI adapters. Libraries were then quantified by qPCR KAPA Biosysytems.		RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina NovaSeq 6000	SRP479683		Lateral_Ectoderm_4.mate1.fastq.gz Lateral_Ectoderm_4.mate2.fastq.gz	fastq fastq	6211000588.0	26929501.0	GSM7989708 r1	0:80.64 1:150	A:1950010947;C:1101223011;G:1497556265;T:1656016608;N:6193757	80	150	1950010947	1101223011	1497556265	1656016608	6193757	SRX23006269	SRS19970080	SRA1774740	Heisenberg group, Institute of Science and Technology Austria (ISTA)	Heisenberg group, Institute of Science and Technology Austria (ISTA)	2	0.65918	0.4633	0.04875	0.02201	0.90623	0.92608	0.79936	0.77887	68	150	B	B	mate2-mate1 similar by mapping diff	illumina	novaseq_era	unknown	cdna_unspecified	nextera	sc	single_cell_plate	smartseq	Austria	2023-12-22	Gastrula	Embryo	Embryo Imprecise	All anatomical structures
29203	29203	SRR27329258	SRX23006268	SRS19970081	SRP479683	PRJNA1055904	BMP dependent patterning of ectoderm tissue material properties modulates lateral mesendoderm cell migration during early zebrafish gastrulation	GSE251904	Transcriptome Analysis	Cell migration is a fundamental process during embryonic development. Most studies in vivo have focussed on the migration of cells using the extracellular matrix ECM as their substrate for migration. In contrast much less is known about how cells migrate on other cells as found in early embryos when the ECM has not yet formed. Here we show that lateral mesendoderm LME cells in the early zebrafish gastrula use the ectoderm as their substrate for migration. We show that the lateral ectoderm is permissive for the animal pole directed migration of LME cells while the ectoderm at the animal pole halts it. These differences in the permissiveness depend on the lateral ectoderm being more cohesive than the animal ectoderm a property controlled by BMP signalling within the ectoderm. Collectively these findings identify ectoderm tissue cohesion as one critical factor regulating LME migration during zebrafish gastrulation. Overall design: To investigate the differences in gene expression between animal and lateral ectoderm causing the difference in material properties between the two ectodermal tissues.	pubmed:40057955	Zebrafish Lateral Ectoderm 7.5 hpf 3	GSM7989707	source name:Lateral Ectoderm\|tissue:Lateral Ectoderm\|genotype:WT TgSebox::eGFP embryos\|treatment:1 cell stage injection of PAmCherry1 mRNA Photoactivation FACS\|geo loc name:missing\|collection date:missing	Zebrafish Lateral Ectoderm 7.5 hpf 3	FastQC quality control Trimmomatic adapter & quality trimming; parameters: 2:30:10 LEADING:3 TRAILING:3 SLIDINGWINDOW:4:15 MINLEN:19 FastQC controlling trimming results salmon Alignment; used flags: seqBias qcBias R DESeq2 differential gene expression analysis; DESEq2 apeglm tximport genefilter GenomicRanges R EnhancedVolcano & biomaRt visualization Assembly: Danio rerio GRCz11 Supplementary files format and content: tab delimited text file with gene name gene length effective length tpm and number of reads for each sample	Lateral Ectoderm	Embryo were injected with 125 pg of PAmCherry1 mRNA at xxx cell stage. At 6 hpf embryos were mounted for up right imaging and a 444 x 220 x 58.5 µm volume was photoactivated using a 405 nm laser either at the animal pole or on the lateral side. The photoactivation process was composed of 10 cycles of 28 sec per embryo in a time window of approximately 90 min. post photoactivation embryos were transferred in Ca2+ free Ringer solution and dissociated by gently pipetting using a P 1000 pipette to obtain a single cell suspension. 5000 PAmCherry1+ cells were sorted in 150 µl of RTL plus lysis buffer RNeasy Plus Micro Kit QIAGEN with 1% of 2 mercaptoethanol.	Cells were subjected to RNA extraction using RNeasy Plus Micro Kit QIAGEN according to the manufacturer instructions Complete cDNA synthesis and library preparation were performed using SMART Seq v3 protocol with Nextera UDI adapters. Libraries were then quantified by qPCR KAPA Biosysytems.	Embryos were kept at 28.5 C until dissociation	tissue:Lateral Ectoderm\|genotype:WT TgSebox::eGFP embryos\|treatment:1 cell stage injection of PAmCherry1 mRNA Photoactivation FACS	GSM7989707	GSM7989707: Zebrafish Lateral Ectoderm 7.5 hpf 3; Danio rerio; RNA Seq	GSM7989707 r1	GSM7989707	1	Cells were subjected to RNA extraction using RNeasy Plus Micro Kit QIAGEN according to the manufacturer instructions Complete cDNA synthesis and library preparation were performed using SMART Seq v3 protocol with Nextera UDI adapters. Libraries were then quantified by qPCR KAPA Biosysytems.		RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina NovaSeq 6000	SRP479683		Lateral_Ectoderm_3.mate1.fastq.gz Lateral_Ectoderm_3.mate2.fastq.gz	fastq fastq	5066130338.0	22014930.0	GSM7989707 r1	0:80.12 1:150	A:1476750755;C:934859886;G:1246347548;T:1403280136;N:4892013	80	150	1476750755	934859886	1246347548	1403280136	4892013	SRX23006268	SRS19970081	SRA1774740	Heisenberg group, Institute of Science and Technology Austria (ISTA)	Heisenberg group, Institute of Science and Technology Austria (ISTA)	2	0.7566	0.52272	0.0462	0.02684	0.89489	0.91246	0.78848	0.77633	68	150	B	B	mate2-mate1 similar by mapping diff	illumina	novaseq_era	unknown	cdna_unspecified	nextera	sc	single_cell_plate	smartseq	Austria	2023-12-22	Gastrula	Embryo	Embryo Imprecise	All anatomical structures
29204	29204	SRR27329259	SRX23006267	SRS19970078	SRP479683	PRJNA1055904	BMP dependent patterning of ectoderm tissue material properties modulates lateral mesendoderm cell migration during early zebrafish gastrulation	GSE251904	Transcriptome Analysis	Cell migration is a fundamental process during embryonic development. Most studies in vivo have focussed on the migration of cells using the extracellular matrix ECM as their substrate for migration. In contrast much less is known about how cells migrate on other cells as found in early embryos when the ECM has not yet formed. Here we show that lateral mesendoderm LME cells in the early zebrafish gastrula use the ectoderm as their substrate for migration. We show that the lateral ectoderm is permissive for the animal pole directed migration of LME cells while the ectoderm at the animal pole halts it. These differences in the permissiveness depend on the lateral ectoderm being more cohesive than the animal ectoderm a property controlled by BMP signalling within the ectoderm. Collectively these findings identify ectoderm tissue cohesion as one critical factor regulating LME migration during zebrafish gastrulation. Overall design: To investigate the differences in gene expression between animal and lateral ectoderm causing the difference in material properties between the two ectodermal tissues.	pubmed:40057955	Zebrafish Lateral Ectoderm 7.5 hpf 2	GSM7989706	source name:Lateral Ectoderm\|tissue:Lateral Ectoderm\|genotype:WT TgSebox::eGFP embryos\|treatment:1 cell stage injection of PAmCherry1 mRNA Photoactivation FACS\|geo loc name:missing\|collection date:missing	Zebrafish Lateral Ectoderm 7.5 hpf 2	FastQC quality control Trimmomatic adapter & quality trimming; parameters: 2:30:10 LEADING:3 TRAILING:3 SLIDINGWINDOW:4:15 MINLEN:19 FastQC controlling trimming results salmon Alignment; used flags: seqBias qcBias R DESeq2 differential gene expression analysis; DESEq2 apeglm tximport genefilter GenomicRanges R EnhancedVolcano & biomaRt visualization Assembly: Danio rerio GRCz11 Supplementary files format and content: tab delimited text file with gene name gene length effective length tpm and number of reads for each sample	Lateral Ectoderm	Embryo were injected with 125 pg of PAmCherry1 mRNA at xxx cell stage. At 6 hpf embryos were mounted for up right imaging and a 444 x 220 x 58.5 µm volume was photoactivated using a 405 nm laser either at the animal pole or on the lateral side. The photoactivation process was composed of 10 cycles of 28 sec per embryo in a time window of approximately 90 min. post photoactivation embryos were transferred in Ca2+ free Ringer solution and dissociated by gently pipetting using a P 1000 pipette to obtain a single cell suspension. 5000 PAmCherry1+ cells were sorted in 150 µl of RTL plus lysis buffer RNeasy Plus Micro Kit QIAGEN with 1% of 2 mercaptoethanol.	Cells were subjected to RNA extraction using RNeasy Plus Micro Kit QIAGEN according to the manufacturer instructions Complete cDNA synthesis and library preparation were performed using SMART Seq v3 protocol with Nextera UDI adapters. Libraries were then quantified by qPCR KAPA Biosysytems.	Embryos were kept at 28.5 C until dissociation	tissue:Lateral Ectoderm\|genotype:WT TgSebox::eGFP embryos\|treatment:1 cell stage injection of PAmCherry1 mRNA Photoactivation FACS	GSM7989706	GSM7989706: Zebrafish Lateral Ectoderm 7.5 hpf 2; Danio rerio; RNA Seq	GSM7989706 r1	GSM7989706	1	Cells were subjected to RNA extraction using RNeasy Plus Micro Kit QIAGEN according to the manufacturer instructions Complete cDNA synthesis and library preparation were performed using SMART Seq v3 protocol with Nextera UDI adapters. Libraries were then quantified by qPCR KAPA Biosysytems.		RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina NovaSeq 6000	SRP479683		Lateral_Ectoderm_2.mate1.fastq.gz Lateral_Ectoderm_2.mate2.fastq.gz	fastq fastq	6372234032.0	27787436.0	GSM7989706 r1	0:79.32 1:150	A:1908123129;C:1195826989;G:1555786583;T:1706205499;N:6291832	79	150	1908123129	1195826989	1555786583	1706205499	6291832	SRX23006267	SRS19970078	SRA1774740	Heisenberg group, Institute of Science and Technology Austria (ISTA)	Heisenberg group, Institute of Science and Technology Austria (ISTA)	2	0.52417	0.50653	0.02922	0.02493	0.89479	0.90938	0.74721	0.74563	68	150	B	B	biological fallback assumption	illumina	novaseq_era	unknown	cdna_unspecified	nextera	sc	single_cell_plate	smartseq	Austria	2023-12-22	Gastrula	Embryo	Embryo Imprecise	All anatomical structures
29205	29205	SRR27329260	SRX23006266	SRS19970079	SRP479683	PRJNA1055904	BMP dependent patterning of ectoderm tissue material properties modulates lateral mesendoderm cell migration during early zebrafish gastrulation	GSE251904	Transcriptome Analysis	Cell migration is a fundamental process during embryonic development. Most studies in vivo have focussed on the migration of cells using the extracellular matrix ECM as their substrate for migration. In contrast much less is known about how cells migrate on other cells as found in early embryos when the ECM has not yet formed. Here we show that lateral mesendoderm LME cells in the early zebrafish gastrula use the ectoderm as their substrate for migration. We show that the lateral ectoderm is permissive for the animal pole directed migration of LME cells while the ectoderm at the animal pole halts it. These differences in the permissiveness depend on the lateral ectoderm being more cohesive than the animal ectoderm a property controlled by BMP signalling within the ectoderm. Collectively these findings identify ectoderm tissue cohesion as one critical factor regulating LME migration during zebrafish gastrulation. Overall design: To investigate the differences in gene expression between animal and lateral ectoderm causing the difference in material properties between the two ectodermal tissues.	pubmed:40057955	Zebrafish Lateral Ectoderm 7.5 hpf 1	GSM7989705	source name:Lateral Ectoderm\|tissue:Lateral Ectoderm\|genotype:WT TgSebox::eGFP embryos\|treatment:1 cell stage injection of PAmCherry1 mRNA Photoactivation FACS\|geo loc name:missing\|collection date:missing	Zebrafish Lateral Ectoderm 7.5 hpf 1	FastQC quality control Trimmomatic adapter & quality trimming; parameters: 2:30:10 LEADING:3 TRAILING:3 SLIDINGWINDOW:4:15 MINLEN:19 FastQC controlling trimming results salmon Alignment; used flags: seqBias qcBias R DESeq2 differential gene expression analysis; DESEq2 apeglm tximport genefilter GenomicRanges R EnhancedVolcano & biomaRt visualization Assembly: Danio rerio GRCz11 Supplementary files format and content: tab delimited text file with gene name gene length effective length tpm and number of reads for each sample	Lateral Ectoderm	Embryo were injected with 125 pg of PAmCherry1 mRNA at xxx cell stage. At 6 hpf embryos were mounted for up right imaging and a 444 x 220 x 58.5 µm volume was photoactivated using a 405 nm laser either at the animal pole or on the lateral side. The photoactivation process was composed of 10 cycles of 28 sec per embryo in a time window of approximately 90 min. post photoactivation embryos were transferred in Ca2+ free Ringer solution and dissociated by gently pipetting using a P 1000 pipette to obtain a single cell suspension. 5000 PAmCherry1+ cells were sorted in 150 µl of RTL plus lysis buffer RNeasy Plus Micro Kit QIAGEN with 1% of 2 mercaptoethanol.	Cells were subjected to RNA extraction using RNeasy Plus Micro Kit QIAGEN according to the manufacturer instructions Complete cDNA synthesis and library preparation were performed using SMART Seq v3 protocol with Nextera UDI adapters. Libraries were then quantified by qPCR KAPA Biosysytems.	Embryos were kept at 28.5 C until dissociation	tissue:Lateral Ectoderm\|genotype:WT TgSebox::eGFP embryos\|treatment:1 cell stage injection of PAmCherry1 mRNA Photoactivation FACS	GSM7989705	GSM7989705: Zebrafish Lateral Ectoderm 7.5 hpf 1; Danio rerio; RNA Seq	GSM7989705 r1	GSM7989705	1	Cells were subjected to RNA extraction using RNeasy Plus Micro Kit QIAGEN according to the manufacturer instructions Complete cDNA synthesis and library preparation were performed using SMART Seq v3 protocol with Nextera UDI adapters. Libraries were then quantified by qPCR KAPA Biosysytems.		RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina NovaSeq 6000	SRP479683		Lateral_Ectoderm_1.mate2.fastq.gz Lateral_Ectoderm_1.mate1.fastq.gz	fastq fastq	10106832046.0	43799441.0	GSM7989705 r1	0:80.75 1:150	A:3384386532;C:1698196797;G:2308026869;T:2706071735;N:10150113	80	150	3384386532	1698196797	2308026869	2706071735	10150113	SRX23006266	SRS19970079	SRA1774740	Heisenberg group, Institute of Science and Technology Austria (ISTA)	Heisenberg group, Institute of Science and Technology Austria (ISTA)	2	0.64457	0.42542	0.04802	0.01964	0.91541	0.93878	0.83455	0.83564	68	150	B	B	mate2-mate1 similar by mapping diff	illumina	novaseq_era	unknown	cdna_unspecified	nextera	sc	single_cell_plate	smartseq	Austria	2023-12-22	Gastrula	Embryo	Embryo Imprecise	All anatomical structures
29206	29206	SRR27329261	SRX23006265	SRS19970077	SRP479683	PRJNA1055904	BMP dependent patterning of ectoderm tissue material properties modulates lateral mesendoderm cell migration during early zebrafish gastrulation	GSE251904	Transcriptome Analysis	Cell migration is a fundamental process during embryonic development. Most studies in vivo have focussed on the migration of cells using the extracellular matrix ECM as their substrate for migration. In contrast much less is known about how cells migrate on other cells as found in early embryos when the ECM has not yet formed. Here we show that lateral mesendoderm LME cells in the early zebrafish gastrula use the ectoderm as their substrate for migration. We show that the lateral ectoderm is permissive for the animal pole directed migration of LME cells while the ectoderm at the animal pole halts it. These differences in the permissiveness depend on the lateral ectoderm being more cohesive than the animal ectoderm a property controlled by BMP signalling within the ectoderm. Collectively these findings identify ectoderm tissue cohesion as one critical factor regulating LME migration during zebrafish gastrulation. Overall design: To investigate the differences in gene expression between animal and lateral ectoderm causing the difference in material properties between the two ectodermal tissues.	pubmed:40057955	Zebrafish Animal Ectoderm 7.5 hpf 4	GSM7989704	source name:Animal Ectoderm\|tissue:Animal Ectoderm\|genotype:WT TgSebox::eGFP embryos\|treatment:1 cell stage injection of PAmCherry1 mRNA Photoactivation FACS\|geo loc name:missing\|collection date:missing	Zebrafish Animal Ectoderm 7.5 hpf 4	FastQC quality control Trimmomatic adapter & quality trimming; parameters: 2:30:10 LEADING:3 TRAILING:3 SLIDINGWINDOW:4:15 MINLEN:19 FastQC controlling trimming results salmon Alignment; used flags: seqBias qcBias R DESeq2 differential gene expression analysis; DESEq2 apeglm tximport genefilter GenomicRanges R EnhancedVolcano & biomaRt visualization Assembly: Danio rerio GRCz11 Supplementary files format and content: tab delimited text file with gene name gene length effective length tpm and number of reads for each sample	Animal Ectoderm	Embryo were injected with 125 pg of PAmCherry1 mRNA at xxx cell stage. At 6 hpf embryos were mounted for up right imaging and a 444 x 220 x 58.5 µm volume was photoactivated using a 405 nm laser either at the animal pole or on the lateral side. The photoactivation process was composed of 10 cycles of 28 sec per embryo in a time window of approximately 90 min. post photoactivation embryos were transferred in Ca2+ free Ringer solution and dissociated by gently pipetting using a P 1000 pipette to obtain a single cell suspension. 5000 PAmCherry1+ cells were sorted in 150 µl of RTL plus lysis buffer RNeasy Plus Micro Kit QIAGEN with 1% of 2 mercaptoethanol.	Cells were subjected to RNA extraction using RNeasy Plus Micro Kit QIAGEN according to the manufacturer instructions Complete cDNA synthesis and library preparation were performed using SMART Seq v3 protocol with Nextera UDI adapters. Libraries were then quantified by qPCR KAPA Biosysytems.	Embryos were kept at 28.5 C until dissociation	tissue:Animal Ectoderm\|genotype:WT TgSebox::eGFP embryos\|treatment:1 cell stage injection of PAmCherry1 mRNA Photoactivation FACS	GSM7989704	GSM7989704: Zebrafish Animal Ectoderm 7.5 hpf 4; Danio rerio; RNA Seq	GSM7989704 r1	GSM7989704	1	Cells were subjected to RNA extraction using RNeasy Plus Micro Kit QIAGEN according to the manufacturer instructions Complete cDNA synthesis and library preparation were performed using SMART Seq v3 protocol with Nextera UDI adapters. Libraries were then quantified by qPCR KAPA Biosysytems.		RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina NovaSeq 6000	SRP479683		Animal_Ectoderm_4.mate1.fastq.gz Animal_Ectoderm_4.mate2.fastq.gz	fastq fastq	5882974802.0	25612144.0	GSM7989704 r1	0:79.69 1:150.00	A:1784390197;C:1112846767;G:1392719296;T:1587227818;N:5790724	79	150	1784390197	1112846767	1392719296	1587227818	5790724	SRX23006265	SRS19970077	SRA1774740	Heisenberg group, Institute of Science and Technology Austria (ISTA)	Heisenberg group, Institute of Science and Technology Austria (ISTA)	2	0.72169	0.51947	0.04749	0.0244	0.89061	0.91007	0.76259	0.75355	90	150	B	B	mate2-mate1 similar by mapping diff	illumina	novaseq_era	unknown	cdna_unspecified	nextera	sc	single_cell_plate	smartseq	Austria	2023-12-22	Gastrula	Embryo	Embryo Imprecise	All anatomical structures
29207	29207	SRR27329262	SRX23006264	SRS19970076	SRP479683	PRJNA1055904	BMP dependent patterning of ectoderm tissue material properties modulates lateral mesendoderm cell migration during early zebrafish gastrulation	GSE251904	Transcriptome Analysis	Cell migration is a fundamental process during embryonic development. Most studies in vivo have focussed on the migration of cells using the extracellular matrix ECM as their substrate for migration. In contrast much less is known about how cells migrate on other cells as found in early embryos when the ECM has not yet formed. Here we show that lateral mesendoderm LME cells in the early zebrafish gastrula use the ectoderm as their substrate for migration. We show that the lateral ectoderm is permissive for the animal pole directed migration of LME cells while the ectoderm at the animal pole halts it. These differences in the permissiveness depend on the lateral ectoderm being more cohesive than the animal ectoderm a property controlled by BMP signalling within the ectoderm. Collectively these findings identify ectoderm tissue cohesion as one critical factor regulating LME migration during zebrafish gastrulation. Overall design: To investigate the differences in gene expression between animal and lateral ectoderm causing the difference in material properties between the two ectodermal tissues.	pubmed:40057955	Zebrafish Animal Ectoderm 7.5 hpf 3	GSM7989703	source name:Animal Ectoderm\|tissue:Animal Ectoderm\|genotype:WT TgSebox::eGFP embryos\|treatment:1 cell stage injection of PAmCherry1 mRNA Photoactivation FACS\|geo loc name:missing\|collection date:missing	Zebrafish Animal Ectoderm 7.5 hpf 3	FastQC quality control Trimmomatic adapter & quality trimming; parameters: 2:30:10 LEADING:3 TRAILING:3 SLIDINGWINDOW:4:15 MINLEN:19 FastQC controlling trimming results salmon Alignment; used flags: seqBias qcBias R DESeq2 differential gene expression analysis; DESEq2 apeglm tximport genefilter GenomicRanges R EnhancedVolcano & biomaRt visualization Assembly: Danio rerio GRCz11 Supplementary files format and content: tab delimited text file with gene name gene length effective length tpm and number of reads for each sample	Animal Ectoderm	Embryo were injected with 125 pg of PAmCherry1 mRNA at xxx cell stage. At 6 hpf embryos were mounted for up right imaging and a 444 x 220 x 58.5 µm volume was photoactivated using a 405 nm laser either at the animal pole or on the lateral side. The photoactivation process was composed of 10 cycles of 28 sec per embryo in a time window of approximately 90 min. post photoactivation embryos were transferred in Ca2+ free Ringer solution and dissociated by gently pipetting using a P 1000 pipette to obtain a single cell suspension. 5000 PAmCherry1+ cells were sorted in 150 µl of RTL plus lysis buffer RNeasy Plus Micro Kit QIAGEN with 1% of 2 mercaptoethanol.	Cells were subjected to RNA extraction using RNeasy Plus Micro Kit QIAGEN according to the manufacturer instructions Complete cDNA synthesis and library preparation were performed using SMART Seq v3 protocol with Nextera UDI adapters. Libraries were then quantified by qPCR KAPA Biosysytems.	Embryos were kept at 28.5 C until dissociation	tissue:Animal Ectoderm\|genotype:WT TgSebox::eGFP embryos\|treatment:1 cell stage injection of PAmCherry1 mRNA Photoactivation FACS	GSM7989703	GSM7989703: Zebrafish Animal Ectoderm 7.5 hpf 3; Danio rerio; RNA Seq	GSM7989703 r1	GSM7989703	1	Cells were subjected to RNA extraction using RNeasy Plus Micro Kit QIAGEN according to the manufacturer instructions Complete cDNA synthesis and library preparation were performed using SMART Seq v3 protocol with Nextera UDI adapters. Libraries were then quantified by qPCR KAPA Biosysytems.		RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina NovaSeq 6000	SRP479683		Animal_Ectoderm_3.mate1.fastq.gz Animal_Ectoderm_3.mate2.fastq.gz	fastq fastq	6802628678.0	29741168.0	GSM7989703 r1	0:78.73 1:150	A:1935418861;C:1362666541;G:1651315180;T:1846520380;N:6707716	78	150	1935418861	1362666541	1651315180	1846520380	6707716	SRX23006264	SRS19970076	SRA1774740	Heisenberg group, Institute of Science and Technology Austria (ISTA)	Heisenberg group, Institute of Science and Technology Austria (ISTA)	2	0.74265	0.48713	0.04983	0.02636	0.87771	0.90335	0.72388	0.71557	68	150	B	B	mate2-mate1 similar by mapping diff	illumina	novaseq_era	unknown	cdna_unspecified	nextera	sc	single_cell_plate	smartseq	Austria	2023-12-22	Gastrula	Embryo	Embryo Imprecise	All anatomical structures
29208	29208	SRR27329263	SRX23006263	SRS19970075	SRP479683	PRJNA1055904	BMP dependent patterning of ectoderm tissue material properties modulates lateral mesendoderm cell migration during early zebrafish gastrulation	GSE251904	Transcriptome Analysis	Cell migration is a fundamental process during embryonic development. Most studies in vivo have focussed on the migration of cells using the extracellular matrix ECM as their substrate for migration. In contrast much less is known about how cells migrate on other cells as found in early embryos when the ECM has not yet formed. Here we show that lateral mesendoderm LME cells in the early zebrafish gastrula use the ectoderm as their substrate for migration. We show that the lateral ectoderm is permissive for the animal pole directed migration of LME cells while the ectoderm at the animal pole halts it. These differences in the permissiveness depend on the lateral ectoderm being more cohesive than the animal ectoderm a property controlled by BMP signalling within the ectoderm. Collectively these findings identify ectoderm tissue cohesion as one critical factor regulating LME migration during zebrafish gastrulation. Overall design: To investigate the differences in gene expression between animal and lateral ectoderm causing the difference in material properties between the two ectodermal tissues.	pubmed:40057955	Zebrafish Animal Ectoderm 7.5 hpf 2	GSM7989702	source name:Animal Ectoderm\|tissue:Animal Ectoderm\|genotype:WT TgSebox::eGFP embryos\|treatment:1 cell stage injection of PAmCherry1 mRNA Photoactivation FACS\|geo loc name:missing\|collection date:missing	Zebrafish Animal Ectoderm 7.5 hpf 2	FastQC quality control Trimmomatic adapter & quality trimming; parameters: 2:30:10 LEADING:3 TRAILING:3 SLIDINGWINDOW:4:15 MINLEN:19 FastQC controlling trimming results salmon Alignment; used flags: seqBias qcBias R DESeq2 differential gene expression analysis; DESEq2 apeglm tximport genefilter GenomicRanges R EnhancedVolcano & biomaRt visualization Assembly: Danio rerio GRCz11 Supplementary files format and content: tab delimited text file with gene name gene length effective length tpm and number of reads for each sample	Animal Ectoderm	Embryo were injected with 125 pg of PAmCherry1 mRNA at xxx cell stage. At 6 hpf embryos were mounted for up right imaging and a 444 x 220 x 58.5 µm volume was photoactivated using a 405 nm laser either at the animal pole or on the lateral side. The photoactivation process was composed of 10 cycles of 28 sec per embryo in a time window of approximately 90 min. post photoactivation embryos were transferred in Ca2+ free Ringer solution and dissociated by gently pipetting using a P 1000 pipette to obtain a single cell suspension. 5000 PAmCherry1+ cells were sorted in 150 µl of RTL plus lysis buffer RNeasy Plus Micro Kit QIAGEN with 1% of 2 mercaptoethanol.	Cells were subjected to RNA extraction using RNeasy Plus Micro Kit QIAGEN according to the manufacturer instructions Complete cDNA synthesis and library preparation were performed using SMART Seq v3 protocol with Nextera UDI adapters. Libraries were then quantified by qPCR KAPA Biosysytems.	Embryos were kept at 28.5 C until dissociation	tissue:Animal Ectoderm\|genotype:WT TgSebox::eGFP embryos\|treatment:1 cell stage injection of PAmCherry1 mRNA Photoactivation FACS	GSM7989702	GSM7989702: Zebrafish Animal Ectoderm 7.5 hpf 2; Danio rerio; RNA Seq	GSM7989702 r1	GSM7989702	1	Cells were subjected to RNA extraction using RNeasy Plus Micro Kit QIAGEN according to the manufacturer instructions Complete cDNA synthesis and library preparation were performed using SMART Seq v3 protocol with Nextera UDI adapters. Libraries were then quantified by qPCR KAPA Biosysytems.		RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina NovaSeq 6000	SRP479683		Animal_Ectoderm_2.mate1.fastq.gz Animal_Ectoderm_2.mate2.fastq.gz	fastq fastq	6083410864.0	26183986.0	GSM7989702 r1	0:82.33 1:150	A:1976903262;C:1110974120;G:1376486325;T:1613990197;N:5056960	82	150	1976903262	1110974120	1376486325	1613990197	5056960	SRX23006263	SRS19970075	SRA1774740	Heisenberg group, Institute of Science and Technology Austria (ISTA)	Heisenberg group, Institute of Science and Technology Austria (ISTA)	2	0.73078	0.55778	0.04052	0.02684	0.89509	0.91106	0.77391	0.7649	68	150	B	B	biological fallback assumption	illumina	novaseq_era	unknown	cdna_unspecified	nextera	sc	single_cell_plate	smartseq	Austria	2023-12-22	Gastrula	Embryo	Embryo Imprecise	All anatomical structures
29209	29209	SRR27329264	SRX23006262	SRS19970074	SRP479683	PRJNA1055904	BMP dependent patterning of ectoderm tissue material properties modulates lateral mesendoderm cell migration during early zebrafish gastrulation	GSE251904	Transcriptome Analysis	Cell migration is a fundamental process during embryonic development. Most studies in vivo have focussed on the migration of cells using the extracellular matrix ECM as their substrate for migration. In contrast much less is known about how cells migrate on other cells as found in early embryos when the ECM has not yet formed. Here we show that lateral mesendoderm LME cells in the early zebrafish gastrula use the ectoderm as their substrate for migration. We show that the lateral ectoderm is permissive for the animal pole directed migration of LME cells while the ectoderm at the animal pole halts it. These differences in the permissiveness depend on the lateral ectoderm being more cohesive than the animal ectoderm a property controlled by BMP signalling within the ectoderm. Collectively these findings identify ectoderm tissue cohesion as one critical factor regulating LME migration during zebrafish gastrulation. Overall design: To investigate the differences in gene expression between animal and lateral ectoderm causing the difference in material properties between the two ectodermal tissues.	pubmed:40057955	Zebrafish Animal Ectoderm 7.5 hpf 1	GSM7989701	source name:Animal Ectoderm\|tissue:Animal Ectoderm\|genotype:WT TgSebox::eGFP embryos\|treatment:1 cell stage injection of PAmCherry1 mRNA Photoactivation FACS\|geo loc name:missing\|collection date:missing	Zebrafish Animal Ectoderm 7.5 hpf 1	FastQC quality control Trimmomatic adapter & quality trimming; parameters: 2:30:10 LEADING:3 TRAILING:3 SLIDINGWINDOW:4:15 MINLEN:19 FastQC controlling trimming results salmon Alignment; used flags: seqBias qcBias R DESeq2 differential gene expression analysis; DESEq2 apeglm tximport genefilter GenomicRanges R EnhancedVolcano & biomaRt visualization Assembly: Danio rerio GRCz11 Supplementary files format and content: tab delimited text file with gene name gene length effective length tpm and number of reads for each sample	Animal Ectoderm	Embryo were injected with 125 pg of PAmCherry1 mRNA at xxx cell stage. At 6 hpf embryos were mounted for up right imaging and a 444 x 220 x 58.5 µm volume was photoactivated using a 405 nm laser either at the animal pole or on the lateral side. The photoactivation process was composed of 10 cycles of 28 sec per embryo in a time window of approximately 90 min. post photoactivation embryos were transferred in Ca2+ free Ringer solution and dissociated by gently pipetting using a P 1000 pipette to obtain a single cell suspension. 5000 PAmCherry1+ cells were sorted in 150 µl of RTL plus lysis buffer RNeasy Plus Micro Kit QIAGEN with 1% of 2 mercaptoethanol.	Cells were subjected to RNA extraction using RNeasy Plus Micro Kit QIAGEN according to the manufacturer instructions Complete cDNA synthesis and library preparation were performed using SMART Seq v3 protocol with Nextera UDI adapters. Libraries were then quantified by qPCR KAPA Biosysytems.	Embryos were kept at 28.5 C until dissociation	tissue:Animal Ectoderm\|genotype:WT TgSebox::eGFP embryos\|treatment:1 cell stage injection of PAmCherry1 mRNA Photoactivation FACS	GSM7989701	GSM7989701: Zebrafish Animal Ectoderm 7.5 hpf 1; Danio rerio; RNA Seq	GSM7989701 r1	GSM7989701	1	Cells were subjected to RNA extraction using RNeasy Plus Micro Kit QIAGEN according to the manufacturer instructions Complete cDNA synthesis and library preparation were performed using SMART Seq v3 protocol with Nextera UDI adapters. Libraries were then quantified by qPCR KAPA Biosysytems.		RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina NovaSeq 6000	SRP479683		Animal_Ectoderm_1.mate1.fastq.gz Animal_Ectoderm_1.mate2.fastq.gz	fastq fastq	7195859200.0	31306406.0	GSM7989701 r1	0:79.85 1:150	A:2025250331;C:1450062111;G:1771726396;T:1941745747;N:7074615	79	150	2025250331	1450062111	1771726396	1941745747	7074615	SRX23006262	SRS19970074	SRA1774740	Heisenberg group, Institute of Science and Technology Austria (ISTA)	Heisenberg group, Institute of Science and Technology Austria (ISTA)	2	0.74201	0.52197	0.04312	0.02595	0.87941	0.9013	0.74298	0.71509	68	150	B	B	mate2-mate1 similar by mapping diff	illumina	novaseq_era	unknown	cdna_unspecified	nextera	sc	single_cell_plate	smartseq	Austria	2023-12-22	Gastrula	Embryo	Embryo Imprecise	All anatomical structures
31773	31773	SRR28590145	SRX24189497	SRS20963520	SRP500320	PRJNA1097627	GPI transamidase complex is required for primordial germ cell migration and development in zebrafish	GSE263481	Transcriptome Analysis	Proteins without xxx domains could be anchored to the cell surface for regulating various biological processes when covalently linked to glycosylphosphatidylinositol GPI molecules by the GPI transamidase GPIT complex. However it remains poorly understood whether and how the GPIT complex affects primordial germ cell PGC development. In this study we report the important roles of GPI transamidase in PGC migration and development in zebrafish embryos. Mutation of pigu or pigk both encoding essential GPIT complex subunits resulted in defective PGC migration with ectopically located PGCs and reduction of PGC counts. Notably a detailed analysis of filopodia in PGCs reveals the attenuated polarity of filopodia distribution along the migration direction in mutant embryos. PGC transplantation and PGC specific rescue experiments demonstrate that both PGC and somatic cell expressed Pigu are required for PGC migration. Furthermore expression levels of PGC specific genes are decreased in pigu mutant PGCs with the derepression of somatic cell genes. Hence we propose that the GPIT complex plays a critical role during PGC migration and development. Overall design: To reveal molecular mechanisms underlying GPI transamidase mediated PGC migration regulation we performed RNA seq of PGCs in wild type and pigu / embryos at the shield and bud stages. We manually picked GFP positive PGCs post cell dissociation of embryos at the shield and bud stages. Then post genotyping of rest somatic cells PGCs with the same genotypes were pooled together and lysed for Smart seq of two biological replicates.	pubmed:39741393	WT shield PGCs RNA rep2	GSM8193374	source name:embryo\|tissue:embryo\|cell line:germ cell\|cell type:PGCs\|genotype:Wile type\|geo loc name:missing\|collection date:missing	WT shield PGCs RNA rep2	RNA seq data were firstly processed using Trim Galore! with default parameters to trim the adapter containing and low quality reads. The filtered data were then mapped to the zebrafish reference genome danRer11 by STAR version: STAR 2.5.3a modified. The gene expression level was normalized to fragments per kilobase of transcript per million mapped FPKM values using Cufflinks version 2.2.1. Assembly: danRer11 Supplementary files format and content: tab delimited ext file includes FPKM values for each samples	embryo	GFP nanos1 3’UTR mRNA was injected into WT and pigu / zygotes respectively to label PGCs. Then the embryos at shield or bud stages were dechorionated and deyolked manually by forceps and transferred to a 1.5 ml Eppendorf tube with 50 µl PBS. Dissections were performed for up to 5 min. The volume of PBS containing embryos was adjusted to 200 µl and then cells were mechanically dissociated by flicking the tube 30 times and pipetting the mixture 10 times through a 214 µl tip. Cells were then transported into a clear Petri dish and GFP labeled PGCs were manually picked under fluorescence microscope with mouth pipette. GFP nanos1 3’UTR mRNA was injected into WT and pigu / zygotes respectively to label PGCs. Then the embryos at shield or bud stages were dechorionated and deyolked manually by forceps and transferred to a 1.5 ml Eppendorf tube with 50 µl PBS. Dissections were performed for up to 5 min. The volume of PBS containing embryos was adjusted to 200 µl and then cells were mechanically dissociated by flicking the tube 30 times and pipetting the mixture 10 times through a 200 µl tip. Cells were then transported into a clear Petri dish and GFP labeled PGCs were manually picked under fluorescence microscope with mouth pipette.	Total RNA of oocytes were purified using RNeasy Mini kit QIAGEN 74104 as recommended protocols. Total RNA 5 μg was DNase I Fermentas EN0521 treated at 37°C for 1 h. Poly A tailed mRNA was collected using DynabeadsTM mRNA Purification Kit Invitrogen 61006. RNA seq was carried out with the SMART seq2 protocol as previously described and subjected to paired end sequencing on the Nextseq 500 Illumina platform according to the manufacturer’s instructions. Notably SMART seq2 of low input RNA seq relies on polyA based amplification.	The wild type Tuebingen Tu and India strains were used throughout the experiments. The transgenic line Tgkop:EGFP nanos 3’UTR citation was described before. Unless stated all embryos were raised in Holtfreter’s solution at 28.5°C and staged as described previously.	tissue:embryo\|cell line:germ cell\|cell type:PGCs\|genotype:Wile type	GSM8193374	GSM8193374: WT shield PGCs RNA rep2; Danio rerio; RNA Seq	GSM8193374 r1	GSM8193374	1	Total RNA of oocytes were purified using RNeasy Mini kit QIAGEN 74104 as recommended protocols. Total RNA 5 μg was DNase I Fermentas EN0521 treated at 37°C for 1 h. Poly A tailed mRNA was collected using DynabeadsTM mRNA Purification Kit Invitrogen 61006. RNA seq was carried out with the SMART seq2 protocol as previously described and subjected to paired end sequencing on the Nextseq 500 Illumina platform according to the manufacturer's instructions. Notably SMART seq2 of low input RNA seq relies on polyA based amplification.		RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	NextSeq 500	SRP500320		WT_shield_PGCs_RNA_rep2_r1.fq.gz WT_shield_PGCs_RNA_rep2_r2.fq.gz	fastq fastq	1476298800.0	4920996.0	GSM8193374 r1	0:150 1:150	A:364072227;C:345274481;G:387289011;T:379660776;N:2305	150	150	364072227	345274481	387289011	379660776	2305	SRX24189497	SRS20963520	SRA1842063	Tsinghua University	Tsinghua University	2	0.9277	0.92845	0.26727	0.2672	0.86803	0.86969	0.75933	0.75985	150	150	B	B	biological fallback assumption	illumina	nextseq	unknown	poly_a	unknown	sc	single_cell_plate	smartseq	China	2024-04-08	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
31774	31774	SRR28590146	SRX24189496	SRS20963519	SRP500320	PRJNA1097627	GPI transamidase complex is required for primordial germ cell migration and development in zebrafish	GSE263481	Transcriptome Analysis	Proteins without xxx domains could be anchored to the cell surface for regulating various biological processes when covalently linked to glycosylphosphatidylinositol GPI molecules by the GPI transamidase GPIT complex. However it remains poorly understood whether and how the GPIT complex affects primordial germ cell PGC development. In this study we report the important roles of GPI transamidase in PGC migration and development in zebrafish embryos. Mutation of pigu or pigk both encoding essential GPIT complex subunits resulted in defective PGC migration with ectopically located PGCs and reduction of PGC counts. Notably a detailed analysis of filopodia in PGCs reveals the attenuated polarity of filopodia distribution along the migration direction in mutant embryos. PGC transplantation and PGC specific rescue experiments demonstrate that both PGC and somatic cell expressed Pigu are required for PGC migration. Furthermore expression levels of PGC specific genes are decreased in pigu mutant PGCs with the derepression of somatic cell genes. Hence we propose that the GPIT complex plays a critical role during PGC migration and development. Overall design: To reveal molecular mechanisms underlying GPI transamidase mediated PGC migration regulation we performed RNA seq of PGCs in wild type and pigu / embryos at the shield and bud stages. We manually picked GFP positive PGCs post cell dissociation of embryos at the shield and bud stages. Then post genotyping of rest somatic cells PGCs with the same genotypes were pooled together and lysed for Smart seq of two biological replicates.	pubmed:39741393	WT shield PGCs RNA rep1	GSM8193373	source name:embryo\|tissue:embryo\|cell line:germ cell\|cell type:PGCs\|genotype:Wile type\|geo loc name:missing\|collection date:missing	WT shield PGCs RNA rep1	RNA seq data were firstly processed using Trim Galore! with default parameters to trim the adapter containing and low quality reads. The filtered data were then mapped to the zebrafish reference genome danRer11 by STAR version: STAR 2.5.3a modified. The gene expression level was normalized to fragments per kilobase of transcript per million mapped FPKM values using Cufflinks version 2.2.1. Assembly: danRer11 Supplementary files format and content: tab delimited ext file includes FPKM values for each samples	embryo	GFP nanos1 3’UTR mRNA was injected into WT and pigu / zygotes respectively to label PGCs. Then the embryos at shield or bud stages were dechorionated and deyolked manually by forceps and transferred to a 1.5 ml Eppendorf tube with 50 µl PBS. Dissections were performed for up to 5 min. The volume of PBS containing embryos was adjusted to 200 µl and then cells were mechanically dissociated by flicking the tube 30 times and pipetting the mixture 10 times through a 212 µl tip. Cells were then transported into a clear Petri dish and GFP labeled PGCs were manually picked under fluorescence microscope with mouth pipette. GFP nanos1 3’UTR mRNA was injected into WT and pigu / zygotes respectively to label PGCs. Then the embryos at shield or bud stages were dechorionated and deyolked manually by forceps and transferred to a 1.5 ml Eppendorf tube with 50 µl PBS. Dissections were performed for up to 5 min. The volume of PBS containing embryos was adjusted to 200 µl and then cells were mechanically dissociated by flicking the tube 30 times and pipetting the mixture 10 times through a 200 µl tip. Cells were then transported into a clear Petri dish and GFP labeled PGCs were manually picked under fluorescence microscope with mouth pipette.	Total RNA of oocytes were purified using RNeasy Mini kit QIAGEN 74104 as recommended protocols. Total RNA 5 μg was DNase I Fermentas EN0521 treated at 37°C for 1 h. Poly A tailed mRNA was collected using DynabeadsTM mRNA Purification Kit Invitrogen 61006. RNA seq was carried out with the SMART seq2 protocol as previously described and subjected to paired end sequencing on the Nextseq 500 Illumina platform according to the manufacturer’s instructions. Notably SMART seq2 of low input RNA seq relies on polyA based amplification.	The wild type Tuebingen Tu and India strains were used throughout the experiments. The transgenic line Tgkop:EGFP nanos 3’UTR citation was described before. Unless stated all embryos were raised in Holtfreter’s solution at 28.5°C and staged as described previously.	tissue:embryo\|cell line:germ cell\|cell type:PGCs\|genotype:Wile type	GSM8193373	GSM8193373: WT shield PGCs RNA rep1; Danio rerio; RNA Seq	GSM8193373 r1	GSM8193373	1	Total RNA of oocytes were purified using RNeasy Mini kit QIAGEN 74104 as recommended protocols. Total RNA 5 μg was DNase I Fermentas EN0521 treated at 37°C for 1 h. Poly A tailed mRNA was collected using DynabeadsTM mRNA Purification Kit Invitrogen 61006. RNA seq was carried out with the SMART seq2 protocol as previously described and subjected to paired end sequencing on the Nextseq 500 Illumina platform according to the manufacturer's instructions. Notably SMART seq2 of low input RNA seq relies on polyA based amplification.		RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	NextSeq 500	SRP500320		WT_shield_PGCs_RNA_rep1_r2.fq.gz WT_shield_PGCs_RNA_rep1_r1.fq.gz	fastq fastq	2758778100.0	9195927.0	GSM8193373 r1	0:150 1:150	A:769058521;C:582931819;G:614181069;T:792602904;N:3787	150	150	769058521	582931819	614181069	792602904	3787	SRX24189496	SRS20963519	SRA1842063	Tsinghua University	Tsinghua University	2	0.93691	0.91581	0.06844	0.06665	0.7683	0.76995	0.53645	0.52363	150	150	B	B	biological fallback assumption	illumina	nextseq	unknown	poly_a	unknown	sc	single_cell_plate	smartseq	China	2024-04-08	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
31775	31775	SRR28590147	SRX24189495	SRS20963518	SRP500320	PRJNA1097627	GPI transamidase complex is required for primordial germ cell migration and development in zebrafish	GSE263481	Transcriptome Analysis	Proteins without xxx domains could be anchored to the cell surface for regulating various biological processes when covalently linked to glycosylphosphatidylinositol GPI molecules by the GPI transamidase GPIT complex. However it remains poorly understood whether and how the GPIT complex affects primordial germ cell PGC development. In this study we report the important roles of GPI transamidase in PGC migration and development in zebrafish embryos. Mutation of pigu or pigk both encoding essential GPIT complex subunits resulted in defective PGC migration with ectopically located PGCs and reduction of PGC counts. Notably a detailed analysis of filopodia in PGCs reveals the attenuated polarity of filopodia distribution along the migration direction in mutant embryos. PGC transplantation and PGC specific rescue experiments demonstrate that both PGC and somatic cell expressed Pigu are required for PGC migration. Furthermore expression levels of PGC specific genes are decreased in pigu mutant PGCs with the derepression of somatic cell genes. Hence we propose that the GPIT complex plays a critical role during PGC migration and development. Overall design: To reveal molecular mechanisms underlying GPI transamidase mediated PGC migration regulation we performed RNA seq of PGCs in wild type and pigu / embryos at the shield and bud stages. We manually picked GFP positive PGCs post cell dissociation of embryos at the shield and bud stages. Then post genotyping of rest somatic cells PGCs with the same genotypes were pooled together and lysed for Smart seq of two biological replicates.	pubmed:39741393	WT bud PGCs RNA rep2	GSM8193372	source name:embryo\|tissue:embryo\|cell line:germ cell\|cell type:PGCs\|genotype:Wile type\|geo loc name:missing\|collection date:missing	WT bud PGCs RNA rep2	RNA seq data were firstly processed using Trim Galore! with default parameters to trim the adapter containing and low quality reads. The filtered data were then mapped to the zebrafish reference genome danRer11 by STAR version: STAR 2.5.3a modified. The gene expression level was normalized to fragments per kilobase of transcript per million mapped FPKM values using Cufflinks version 2.2.1. Assembly: danRer11 Supplementary files format and content: tab delimited ext file includes FPKM values for each samples	embryo	GFP nanos1 3’UTR mRNA was injected into WT and pigu / zygotes respectively to label PGCs. Then the embryos at shield or bud stages were dechorionated and deyolked manually by forceps and transferred to a 1.5 ml Eppendorf tube with 50 µl PBS. Dissections were performed for up to 5 min. The volume of PBS containing embryos was adjusted to 200 µl and then cells were mechanically dissociated by flicking the tube 30 times and pipetting the mixture 10 times through a 210 µl tip. Cells were then transported into a clear Petri dish and GFP labeled PGCs were manually picked under fluorescence microscope with mouth pipette. GFP nanos1 3’UTR mRNA was injected into WT and pigu / zygotes respectively to label PGCs. Then the embryos at shield or bud stages were dechorionated and deyolked manually by forceps and transferred to a 1.5 ml Eppendorf tube with 50 µl PBS. Dissections were performed for up to 5 min. The volume of PBS containing embryos was adjusted to 200 µl and then cells were mechanically dissociated by flicking the tube 30 times and pipetting the mixture 10 times through a 200 µl tip. Cells were then transported into a clear Petri dish and GFP labeled PGCs were manually picked under fluorescence microscope with mouth pipette.	Total RNA of oocytes were purified using RNeasy Mini kit QIAGEN 74104 as recommended protocols. Total RNA 5 μg was DNase I Fermentas EN0521 treated at 37°C for 1 h. Poly A tailed mRNA was collected using DynabeadsTM mRNA Purification Kit Invitrogen 61006. RNA seq was carried out with the SMART seq2 protocol as previously described and subjected to paired end sequencing on the Nextseq 500 Illumina platform according to the manufacturer’s instructions. Notably SMART seq2 of low input RNA seq relies on polyA based amplification.	The wild type Tuebingen Tu and India strains were used throughout the experiments. The transgenic line Tgkop:EGFP nanos 3’UTR citation was described before. Unless stated all embryos were raised in Holtfreter’s solution at 28.5°C and staged as described previously.	tissue:embryo\|cell line:germ cell\|cell type:PGCs\|genotype:Wile type	GSM8193372	GSM8193372: WT bud PGCs RNA rep2; Danio rerio; RNA Seq	GSM8193372 r1	GSM8193372	1	Total RNA of oocytes were purified using RNeasy Mini kit QIAGEN 74104 as recommended protocols. Total RNA 5 μg was DNase I Fermentas EN0521 treated at 37°C for 1 h. Poly A tailed mRNA was collected using DynabeadsTM mRNA Purification Kit Invitrogen 61006. RNA seq was carried out with the SMART seq2 protocol as previously described and subjected to paired end sequencing on the Nextseq 500 Illumina platform according to the manufacturer's instructions. Notably SMART seq2 of low input RNA seq relies on polyA based amplification.		RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	NextSeq 500	SRP500320		WT_bud_PGCs_RNA_rep2_r1.fq.gz WT_bud_PGCs_RNA_rep2_r2.fq.gz	fastq fastq	3111261300.0	10370871.0	GSM8193372 r1	0:150 1:150	A:817262319;C:691724200;G:751270343;T:850999265;N:5173	150	150	817262319	691724200	751270343	850999265	5173	SRX24189495	SRS20963518	SRA1842063	Tsinghua University	Tsinghua University	2	0.90729	0.9053	0.11596	0.1166	0.79963	0.79981	0.60276	0.59912	150	150	B	B	biological fallback assumption	illumina	nextseq	unknown	poly_a	unknown	sc	single_cell_plate	smartseq	China	2024-04-08	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
31776	31776	SRR28590148	SRX24189494	SRS20963517	SRP500320	PRJNA1097627	GPI transamidase complex is required for primordial germ cell migration and development in zebrafish	GSE263481	Transcriptome Analysis	Proteins without xxx domains could be anchored to the cell surface for regulating various biological processes when covalently linked to glycosylphosphatidylinositol GPI molecules by the GPI transamidase GPIT complex. However it remains poorly understood whether and how the GPIT complex affects primordial germ cell PGC development. In this study we report the important roles of GPI transamidase in PGC migration and development in zebrafish embryos. Mutation of pigu or pigk both encoding essential GPIT complex subunits resulted in defective PGC migration with ectopically located PGCs and reduction of PGC counts. Notably a detailed analysis of filopodia in PGCs reveals the attenuated polarity of filopodia distribution along the migration direction in mutant embryos. PGC transplantation and PGC specific rescue experiments demonstrate that both PGC and somatic cell expressed Pigu are required for PGC migration. Furthermore expression levels of PGC specific genes are decreased in pigu mutant PGCs with the derepression of somatic cell genes. Hence we propose that the GPIT complex plays a critical role during PGC migration and development. Overall design: To reveal molecular mechanisms underlying GPI transamidase mediated PGC migration regulation we performed RNA seq of PGCs in wild type and pigu / embryos at the shield and bud stages. We manually picked GFP positive PGCs post cell dissociation of embryos at the shield and bud stages. Then post genotyping of rest somatic cells PGCs with the same genotypes were pooled together and lysed for Smart seq of two biological replicates.	pubmed:39741393	WT bud PGCs RNA rep1	GSM8193371	source name:embryo\|tissue:embryo\|cell line:germ cell\|cell type:PGCs\|genotype:Wile type\|geo loc name:missing\|collection date:missing	WT bud PGCs RNA rep1	RNA seq data were firstly processed using Trim Galore! with default parameters to trim the adapter containing and low quality reads. The filtered data were then mapped to the zebrafish reference genome danRer11 by STAR version: STAR 2.5.3a modified. The gene expression level was normalized to fragments per kilobase of transcript per million mapped FPKM values using Cufflinks version 2.2.1. Assembly: danRer11 Supplementary files format and content: tab delimited ext file includes FPKM values for each samples	embryo	GFP nanos1 3’UTR mRNA was injected into WT and pigu / zygotes respectively to label PGCs. Then the embryos at shield or bud stages were dechorionated and deyolked manually by forceps and transferred to a 1.5 ml Eppendorf tube with 50 µl PBS. Dissections were performed for up to 5 min. The volume of PBS containing embryos was adjusted to 200 µl and then cells were mechanically dissociated by flicking the tube 30 times and pipetting the mixture 10 times through a 208 µl tip. Cells were then transported into a clear Petri dish and GFP labeled PGCs were manually picked under fluorescence microscope with mouth pipette. GFP nanos1 3’UTR mRNA was injected into WT and pigu / zygotes respectively to label PGCs. Then the embryos at shield or bud stages were dechorionated and deyolked manually by forceps and transferred to a 1.5 ml Eppendorf tube with 50 µl PBS. Dissections were performed for up to 5 min. The volume of PBS containing embryos was adjusted to 200 µl and then cells were mechanically dissociated by flicking the tube 30 times and pipetting the mixture 10 times through a 200 µl tip. Cells were then transported into a clear Petri dish and GFP labeled PGCs were manually picked under fluorescence microscope with mouth pipette.	Total RNA of oocytes were purified using RNeasy Mini kit QIAGEN 74104 as recommended protocols. Total RNA 5 μg was DNase I Fermentas EN0521 treated at 37°C for 1 h. Poly A tailed mRNA was collected using DynabeadsTM mRNA Purification Kit Invitrogen 61006. RNA seq was carried out with the SMART seq2 protocol as previously described and subjected to paired end sequencing on the Nextseq 500 Illumina platform according to the manufacturer’s instructions. Notably SMART seq2 of low input RNA seq relies on polyA based amplification.	The wild type Tuebingen Tu and India strains were used throughout the experiments. The transgenic line Tgkop:EGFP nanos 3’UTR citation was described before. Unless stated all embryos were raised in Holtfreter’s solution at 28.5°C and staged as described previously.	tissue:embryo\|cell line:germ cell\|cell type:PGCs\|genotype:Wile type	GSM8193371	GSM8193371: WT bud PGCs RNA rep1; Danio rerio; RNA Seq	GSM8193371 r1	GSM8193371	1	Total RNA of oocytes were purified using RNeasy Mini kit QIAGEN 74104 as recommended protocols. Total RNA 5 μg was DNase I Fermentas EN0521 treated at 37°C for 1 h. Poly A tailed mRNA was collected using DynabeadsTM mRNA Purification Kit Invitrogen 61006. RNA seq was carried out with the SMART seq2 protocol as previously described and subjected to paired end sequencing on the Nextseq 500 Illumina platform according to the manufacturer's instructions. Notably SMART seq2 of low input RNA seq relies on polyA based amplification.		RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	NextSeq 500	SRP500320		WT_bud_PGCs_RNA_rep1_r1.fq.gz WT_bud_PGCs_RNA_rep1_r2.fq.gz	fastq fastq	3758269200.0	12527564.0	GSM8193371 r1	0:150 1:150	A:1041649283;C:801966468;G:845188264;T:1069459832;N:5353	150	150	1041649283	801966468	845188264	1069459832	5353	SRX24189494	SRS20963517	SRA1842063	Tsinghua University	Tsinghua University	2	0.93714	0.91161	0.07974	0.07673	0.75785	0.76037	0.52906	0.52822	150	150	B	B	biological fallback assumption	illumina	nextseq	unknown	poly_a	unknown	sc	single_cell_plate	smartseq	China	2024-04-08	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
31777	31777	SRR28590149	SRX24189493	SRS20963516	SRP500320	PRJNA1097627	GPI transamidase complex is required for primordial germ cell migration and development in zebrafish	GSE263481	Transcriptome Analysis	Proteins without xxx domains could be anchored to the cell surface for regulating various biological processes when covalently linked to glycosylphosphatidylinositol GPI molecules by the GPI transamidase GPIT complex. However it remains poorly understood whether and how the GPIT complex affects primordial germ cell PGC development. In this study we report the important roles of GPI transamidase in PGC migration and development in zebrafish embryos. Mutation of pigu or pigk both encoding essential GPIT complex subunits resulted in defective PGC migration with ectopically located PGCs and reduction of PGC counts. Notably a detailed analysis of filopodia in PGCs reveals the attenuated polarity of filopodia distribution along the migration direction in mutant embryos. PGC transplantation and PGC specific rescue experiments demonstrate that both PGC and somatic cell expressed Pigu are required for PGC migration. Furthermore expression levels of PGC specific genes are decreased in pigu mutant PGCs with the derepression of somatic cell genes. Hence we propose that the GPIT complex plays a critical role during PGC migration and development. Overall design: To reveal molecular mechanisms underlying GPI transamidase mediated PGC migration regulation we performed RNA seq of PGCs in wild type and pigu / embryos at the shield and bud stages. We manually picked GFP positive PGCs post cell dissociation of embryos at the shield and bud stages. Then post genotyping of rest somatic cells PGCs with the same genotypes were pooled together and lysed for Smart seq of two biological replicates.	pubmed:39741393	pigu homo shield PGCs RNA rep2	GSM8193370	source name:embryo\|tissue:embryo\|cell line:germ cell\|cell type:PGCs\|genotype:pigu homo\|geo loc name:missing\|collection date:missing	pigu homo shield PGCs RNA rep2	RNA seq data were firstly processed using Trim Galore! with default parameters to trim the adapter containing and low quality reads. The filtered data were then mapped to the zebrafish reference genome danRer11 by STAR version: STAR 2.5.3a modified. The gene expression level was normalized to fragments per kilobase of transcript per million mapped FPKM values using Cufflinks version 2.2.1. Assembly: danRer11 Supplementary files format and content: tab delimited ext file includes FPKM values for each samples	embryo	GFP nanos1 3’UTR mRNA was injected into WT and pigu / zygotes respectively to label PGCs. Then the embryos at shield or bud stages were dechorionated and deyolked manually by forceps and transferred to a 1.5 ml Eppendorf tube with 50 µl PBS. Dissections were performed for up to 5 min. The volume of PBS containing embryos was adjusted to 200 µl and then cells were mechanically dissociated by flicking the tube 30 times and pipetting the mixture 10 times through a 206 µl tip. Cells were then transported into a clear Petri dish and GFP labeled PGCs were manually picked under fluorescence microscope with mouth pipette. GFP nanos1 3’UTR mRNA was injected into WT and pigu / zygotes respectively to label PGCs. Then the embryos at shield or bud stages were dechorionated and deyolked manually by forceps and transferred to a 1.5 ml Eppendorf tube with 50 µl PBS. Dissections were performed for up to 5 min. The volume of PBS containing embryos was adjusted to 200 µl and then cells were mechanically dissociated by flicking the tube 30 times and pipetting the mixture 10 times through a 200 µl tip. Cells were then transported into a clear Petri dish and GFP labeled PGCs were manually picked under fluorescence microscope with mouth pipette.	Total RNA of oocytes were purified using RNeasy Mini kit QIAGEN 74104 as recommended protocols. Total RNA 5 μg was DNase I Fermentas EN0521 treated at 37°C for 1 h. Poly A tailed mRNA was collected using DynabeadsTM mRNA Purification Kit Invitrogen 61006. RNA seq was carried out with the SMART seq2 protocol as previously described and subjected to paired end sequencing on the Nextseq 500 Illumina platform according to the manufacturer’s instructions. Notably SMART seq2 of low input RNA seq relies on polyA based amplification.	The wild type Tuebingen Tu and India strains were used throughout the experiments. The transgenic line Tgkop:EGFP nanos 3’UTR citation was described before. Unless stated all embryos were raised in Holtfreter’s solution at 28.5°C and staged as described previously.	tissue:embryo\|cell line:germ cell\|cell type:PGCs\|genotype:pigu homo	GSM8193370	GSM8193370: pigu homo shield PGCs RNA rep2; Danio rerio; RNA Seq	GSM8193370 r1	GSM8193370	1	Total RNA of oocytes were purified using RNeasy Mini kit QIAGEN 74104 as recommended protocols. Total RNA 5 μg was DNase I Fermentas EN0521 treated at 37°C for 1 h. Poly A tailed mRNA was collected using DynabeadsTM mRNA Purification Kit Invitrogen 61006. RNA seq was carried out with the SMART seq2 protocol as previously described and subjected to paired end sequencing on the Nextseq 500 Illumina platform according to the manufacturer's instructions. Notably SMART seq2 of low input RNA seq relies on polyA based amplification.		RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	NextSeq 500	SRP500320		pigu_homo_shield_PGCs_RNA_rep2_r2.fq.gz pigu_homo_shield_PGCs_RNA_rep2_r1.fq.gz	fastq fastq	875630700.0	2918769.0	GSM8193370 r1	0:150 1:150	A:229625544;C:190227589;G:217096984;T:238679075;N:1508	150	150	229625544	190227589	217096984	238679075	1508	SRX24189493	SRS20963516	SRA1842063	Tsinghua University	Tsinghua University	2	0.81852	0.81643	0.19346	0.19329	0.89852	0.89897	0.83525	0.83082	150	150	B	B	biological fallback assumption	illumina	nextseq	unknown	poly_a	unknown	sc	single_cell_plate	smartseq	China	2024-04-08	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
31778	31778	SRR28590150	SRX24189492	SRS20963515	SRP500320	PRJNA1097627	GPI transamidase complex is required for primordial germ cell migration and development in zebrafish	GSE263481	Transcriptome Analysis	Proteins without xxx domains could be anchored to the cell surface for regulating various biological processes when covalently linked to glycosylphosphatidylinositol GPI molecules by the GPI transamidase GPIT complex. However it remains poorly understood whether and how the GPIT complex affects primordial germ cell PGC development. In this study we report the important roles of GPI transamidase in PGC migration and development in zebrafish embryos. Mutation of pigu or pigk both encoding essential GPIT complex subunits resulted in defective PGC migration with ectopically located PGCs and reduction of PGC counts. Notably a detailed analysis of filopodia in PGCs reveals the attenuated polarity of filopodia distribution along the migration direction in mutant embryos. PGC transplantation and PGC specific rescue experiments demonstrate that both PGC and somatic cell expressed Pigu are required for PGC migration. Furthermore expression levels of PGC specific genes are decreased in pigu mutant PGCs with the derepression of somatic cell genes. Hence we propose that the GPIT complex plays a critical role during PGC migration and development. Overall design: To reveal molecular mechanisms underlying GPI transamidase mediated PGC migration regulation we performed RNA seq of PGCs in wild type and pigu / embryos at the shield and bud stages. We manually picked GFP positive PGCs post cell dissociation of embryos at the shield and bud stages. Then post genotyping of rest somatic cells PGCs with the same genotypes were pooled together and lysed for Smart seq of two biological replicates.	pubmed:39741393	pigu homo shield PGCs RNA rep1	GSM8193369	source name:embryo\|tissue:embryo\|cell line:germ cell\|cell type:PGCs\|genotype:pigu homo\|geo loc name:missing\|collection date:missing	pigu homo shield PGCs RNA rep1	RNA seq data were firstly processed using Trim Galore! with default parameters to trim the adapter containing and low quality reads. The filtered data were then mapped to the zebrafish reference genome danRer11 by STAR version: STAR 2.5.3a modified. The gene expression level was normalized to fragments per kilobase of transcript per million mapped FPKM values using Cufflinks version 2.2.1. Assembly: danRer11 Supplementary files format and content: tab delimited ext file includes FPKM values for each samples	embryo	GFP nanos1 3’UTR mRNA was injected into WT and pigu / zygotes respectively to label PGCs. Then the embryos at shield or bud stages were dechorionated and deyolked manually by forceps and transferred to a 1.5 ml Eppendorf tube with 50 µl PBS. Dissections were performed for up to 5 min. The volume of PBS containing embryos was adjusted to 200 µl and then cells were mechanically dissociated by flicking the tube 30 times and pipetting the mixture 10 times through a 204 µl tip. Cells were then transported into a clear Petri dish and GFP labeled PGCs were manually picked under fluorescence microscope with mouth pipette. GFP nanos1 3’UTR mRNA was injected into WT and pigu / zygotes respectively to label PGCs. Then the embryos at shield or bud stages were dechorionated and deyolked manually by forceps and transferred to a 1.5 ml Eppendorf tube with 50 µl PBS. Dissections were performed for up to 5 min. The volume of PBS containing embryos was adjusted to 200 µl and then cells were mechanically dissociated by flicking the tube 30 times and pipetting the mixture 10 times through a 200 µl tip. Cells were then transported into a clear Petri dish and GFP labeled PGCs were manually picked under fluorescence microscope with mouth pipette.	Total RNA of oocytes were purified using RNeasy Mini kit QIAGEN 74104 as recommended protocols. Total RNA 5 μg was DNase I Fermentas EN0521 treated at 37°C for 1 h. Poly A tailed mRNA was collected using DynabeadsTM mRNA Purification Kit Invitrogen 61006. RNA seq was carried out with the SMART seq2 protocol as previously described and subjected to paired end sequencing on the Nextseq 500 Illumina platform according to the manufacturer’s instructions. Notably SMART seq2 of low input RNA seq relies on polyA based amplification.	The wild type Tuebingen Tu and India strains were used throughout the experiments. The transgenic line Tgkop:EGFP nanos 3’UTR citation was described before. Unless stated all embryos were raised in Holtfreter’s solution at 28.5°C and staged as described previously.	tissue:embryo\|cell line:germ cell\|cell type:PGCs\|genotype:pigu homo	GSM8193369	GSM8193369: pigu homo shield PGCs RNA rep1; Danio rerio; RNA Seq	GSM8193369 r1	GSM8193369	1	Total RNA of oocytes were purified using RNeasy Mini kit QIAGEN 74104 as recommended protocols. Total RNA 5 μg was DNase I Fermentas EN0521 treated at 37°C for 1 h. Poly A tailed mRNA was collected using DynabeadsTM mRNA Purification Kit Invitrogen 61006. RNA seq was carried out with the SMART seq2 protocol as previously described and subjected to paired end sequencing on the Nextseq 500 Illumina platform according to the manufacturer's instructions. Notably SMART seq2 of low input RNA seq relies on polyA based amplification.		RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	NextSeq 500	SRP500320		pigu_homo_shield_PGCs_RNA_rep1_r2.fq.gz pigu_homo_shield_PGCs_RNA_rep1_r1.fq.gz	fastq fastq	2204383800.0	7347946.0	GSM8193369 r1	0:150 1:150	A:618947181;C:459789570;G:491705141;T:633938869;N:3039	150	150	618947181	459789570	491705141	633938869	3039	SRX24189492	SRS20963515	SRA1842063	Tsinghua University	Tsinghua University	2	0.91971	0.9203	0.0732	0.07256	0.77512	0.77516	0.53347	0.53419	150	150	B	B	biological fallback assumption	illumina	nextseq	unknown	poly_a	unknown	sc	single_cell_plate	smartseq	China	2024-04-08	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
31779	31779	SRR28590151	SRX24189491	SRS20963514	SRP500320	PRJNA1097627	GPI transamidase complex is required for primordial germ cell migration and development in zebrafish	GSE263481	Transcriptome Analysis	Proteins without xxx domains could be anchored to the cell surface for regulating various biological processes when covalently linked to glycosylphosphatidylinositol GPI molecules by the GPI transamidase GPIT complex. However it remains poorly understood whether and how the GPIT complex affects primordial germ cell PGC development. In this study we report the important roles of GPI transamidase in PGC migration and development in zebrafish embryos. Mutation of pigu or pigk both encoding essential GPIT complex subunits resulted in defective PGC migration with ectopically located PGCs and reduction of PGC counts. Notably a detailed analysis of filopodia in PGCs reveals the attenuated polarity of filopodia distribution along the migration direction in mutant embryos. PGC transplantation and PGC specific rescue experiments demonstrate that both PGC and somatic cell expressed Pigu are required for PGC migration. Furthermore expression levels of PGC specific genes are decreased in pigu mutant PGCs with the derepression of somatic cell genes. Hence we propose that the GPIT complex plays a critical role during PGC migration and development. Overall design: To reveal molecular mechanisms underlying GPI transamidase mediated PGC migration regulation we performed RNA seq of PGCs in wild type and pigu / embryos at the shield and bud stages. We manually picked GFP positive PGCs post cell dissociation of embryos at the shield and bud stages. Then post genotyping of rest somatic cells PGCs with the same genotypes were pooled together and lysed for Smart seq of two biological replicates.	pubmed:39741393	pigu homo bud PGCs RNA rep2	GSM8193368	source name:embryo\|tissue:embryo\|cell line:germ cell\|cell type:PGCs\|genotype:pigu homo\|geo loc name:missing\|collection date:missing	pigu homo bud PGCs RNA rep2	RNA seq data were firstly processed using Trim Galore! with default parameters to trim the adapter containing and low quality reads. The filtered data were then mapped to the zebrafish reference genome danRer11 by STAR version: STAR 2.5.3a modified. The gene expression level was normalized to fragments per kilobase of transcript per million mapped FPKM values using Cufflinks version 2.2.1. Assembly: danRer11 Supplementary files format and content: tab delimited ext file includes FPKM values for each samples	embryo	GFP nanos1 3’UTR mRNA was injected into WT and pigu / zygotes respectively to label PGCs. Then the embryos at shield or bud stages were dechorionated and deyolked manually by forceps and transferred to a 1.5 ml Eppendorf tube with 50 µl PBS. Dissections were performed for up to 5 min. The volume of PBS containing embryos was adjusted to 200 µl and then cells were mechanically dissociated by flicking the tube 30 times and pipetting the mixture 10 times through a 202 µl tip. Cells were then transported into a clear Petri dish and GFP labeled PGCs were manually picked under fluorescence microscope with mouth pipette. GFP nanos1 3’UTR mRNA was injected into WT and pigu / zygotes respectively to label PGCs. Then the embryos at shield or bud stages were dechorionated and deyolked manually by forceps and transferred to a 1.5 ml Eppendorf tube with 50 µl PBS. Dissections were performed for up to 5 min. The volume of PBS containing embryos was adjusted to 200 µl and then cells were mechanically dissociated by flicking the tube 30 times and pipetting the mixture 10 times through a 200 µl tip. Cells were then transported into a clear Petri dish and GFP labeled PGCs were manually picked under fluorescence microscope with mouth pipette.	Total RNA of oocytes were purified using RNeasy Mini kit QIAGEN 74104 as recommended protocols. Total RNA 5 μg was DNase I Fermentas EN0521 treated at 37°C for 1 h. Poly A tailed mRNA was collected using DynabeadsTM mRNA Purification Kit Invitrogen 61006. RNA seq was carried out with the SMART seq2 protocol as previously described and subjected to paired end sequencing on the Nextseq 500 Illumina platform according to the manufacturer’s instructions. Notably SMART seq2 of low input RNA seq relies on polyA based amplification.	The wild type Tuebingen Tu and India strains were used throughout the experiments. The transgenic line Tgkop:EGFP nanos 3’UTR citation was described before. Unless stated all embryos were raised in Holtfreter’s solution at 28.5°C and staged as described previously.	tissue:embryo\|cell line:germ cell\|cell type:PGCs\|genotype:pigu homo	GSM8193368	GSM8193368: pigu homo bud PGCs RNA rep2; Danio rerio; RNA Seq	GSM8193368 r1	GSM8193368	1	Total RNA of oocytes were purified using RNeasy Mini kit QIAGEN 74104 as recommended protocols. Total RNA 5 μg was DNase I Fermentas EN0521 treated at 37°C for 1 h. Poly A tailed mRNA was collected using DynabeadsTM mRNA Purification Kit Invitrogen 61006. RNA seq was carried out with the SMART seq2 protocol as previously described and subjected to paired end sequencing on the Nextseq 500 Illumina platform according to the manufacturer's instructions. Notably SMART seq2 of low input RNA seq relies on polyA based amplification.		RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	NextSeq 500	SRP500320		pigu_homo_bud_PGCs_RNA_rep2_r1.fq.gz pigu_homo_bud_PGCs_RNA_rep2_r2.fq.gz	fastq fastq	2137638000.0	7125460.0	GSM8193368 r1	0:150 1:150	A:588045273;C:460739300;G:496293927;T:592556088;N:3412	150	150	588045273	460739300	496293927	592556088	3412	SRX24189491	SRS20963514	SRA1842063	Tsinghua University	Tsinghua University	2	0.36533	0.36278	0.05703	0.057	0.87754	0.87811	0.60206	0.59677	150	150	B	B	biological fallback assumption	illumina	nextseq	unknown	poly_a	unknown	sc	single_cell_plate	smartseq	China	2024-04-08	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
31780	31780	SRR28590152	SRX24189490	SRS20963513	SRP500320	PRJNA1097627	GPI transamidase complex is required for primordial germ cell migration and development in zebrafish	GSE263481	Transcriptome Analysis	Proteins without xxx domains could be anchored to the cell surface for regulating various biological processes when covalently linked to glycosylphosphatidylinositol GPI molecules by the GPI transamidase GPIT complex. However it remains poorly understood whether and how the GPIT complex affects primordial germ cell PGC development. In this study we report the important roles of GPI transamidase in PGC migration and development in zebrafish embryos. Mutation of pigu or pigk both encoding essential GPIT complex subunits resulted in defective PGC migration with ectopically located PGCs and reduction of PGC counts. Notably a detailed analysis of filopodia in PGCs reveals the attenuated polarity of filopodia distribution along the migration direction in mutant embryos. PGC transplantation and PGC specific rescue experiments demonstrate that both PGC and somatic cell expressed Pigu are required for PGC migration. Furthermore expression levels of PGC specific genes are decreased in pigu mutant PGCs with the derepression of somatic cell genes. Hence we propose that the GPIT complex plays a critical role during PGC migration and development. Overall design: To reveal molecular mechanisms underlying GPI transamidase mediated PGC migration regulation we performed RNA seq of PGCs in wild type and pigu / embryos at the shield and bud stages. We manually picked GFP positive PGCs post cell dissociation of embryos at the shield and bud stages. Then post genotyping of rest somatic cells PGCs with the same genotypes were pooled together and lysed for Smart seq of two biological replicates.	pubmed:39741393	pigu homo bud PGCs RNA rep1	GSM8193367	source name:embryo\|tissue:embryo\|cell line:germ cell\|cell type:PGCs\|genotype:pigu homo\|geo loc name:missing\|collection date:missing	pigu homo bud PGCs RNA rep1	RNA seq data were firstly processed using Trim Galore! with default parameters to trim the adapter containing and low quality reads. The filtered data were then mapped to the zebrafish reference genome danRer11 by STAR version: STAR 2.5.3a modified. The gene expression level was normalized to fragments per kilobase of transcript per million mapped FPKM values using Cufflinks version 2.2.1. Assembly: danRer11 Supplementary files format and content: tab delimited ext file includes FPKM values for each samples	embryo	GFP nanos1 3’UTR mRNA was injected into WT and pigu / zygotes respectively to label PGCs. Then the embryos at shield or bud stages were dechorionated and deyolked manually by forceps and transferred to a 1.5 ml Eppendorf tube with 50 µl PBS. Dissections were performed for up to 5 min. The volume of PBS containing embryos was adjusted to 200 µl and then cells were mechanically dissociated by flicking the tube 30 times and pipetting the mixture 10 times through a 200 µl tip. Cells were then transported into a clear Petri dish and GFP labeled PGCs were manually picked under fluorescence microscope with mouth pipette. GFP nanos1 3’UTR mRNA was injected into WT and pigu / zygotes respectively to label PGCs. Then the embryos at shield or bud stages were dechorionated and deyolked manually by forceps and transferred to a 1.5 ml Eppendorf tube with 50 µl PBS. Dissections were performed for up to 5 min. The volume of PBS containing embryos was adjusted to 200 µl and then cells were mechanically dissociated by flicking the tube 30 times and pipetting the mixture 10 times through a 200 µl tip. Cells were then transported into a clear Petri dish and GFP labeled PGCs were manually picked under fluorescence microscope with mouth pipette.	Total RNA of oocytes were purified using RNeasy Mini kit QIAGEN 74104 as recommended protocols. Total RNA 5 μg was DNase I Fermentas EN0521 treated at 37°C for 1 h. Poly A tailed mRNA was collected using DynabeadsTM mRNA Purification Kit Invitrogen 61006. RNA seq was carried out with the SMART seq2 protocol as previously described and subjected to paired end sequencing on the Nextseq 500 Illumina platform according to the manufacturer’s instructions. Notably SMART seq2 of low input RNA seq relies on polyA based amplification.	The wild type Tuebingen Tu and India strains were used throughout the experiments. The transgenic line Tgkop:EGFP nanos 3’UTR citation was described before. Unless stated all embryos were raised in Holtfreter’s solution at 28.5°C and staged as described previously.	tissue:embryo\|cell line:germ cell\|cell type:PGCs\|genotype:pigu homo	GSM8193367	GSM8193367: pigu homo bud PGCs RNA rep1; Danio rerio; RNA Seq	GSM8193367 r1	GSM8193367	1	Total RNA of oocytes were purified using RNeasy Mini kit QIAGEN 74104 as recommended protocols. Total RNA 5 μg was DNase I Fermentas EN0521 treated at 37°C for 1 h. Poly A tailed mRNA was collected using DynabeadsTM mRNA Purification Kit Invitrogen 61006. RNA seq was carried out with the SMART seq2 protocol as previously described and subjected to paired end sequencing on the Nextseq 500 Illumina platform according to the manufacturer's instructions. Notably SMART seq2 of low input RNA seq relies on polyA based amplification.		RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	NextSeq 500	SRP500320		pigu_homo_bud_PGCs_RNA_rep1_r1.fq.gz pigu_homo_bud_PGCs_RNA_rep1_r2.fq.gz	fastq fastq	2734187100.0	9113957.0	GSM8193367 r1	0:150 1:150	A:729072634;C:614811431;G:644751435;T:745548063;N:3537	150	150	729072634	614811431	644751435	745548063	3537	SRX24189490	SRS20963513	SRA1842063	Tsinghua University	Tsinghua University	2	0.94018	0.93763	0.11685	0.1143	0.77749	0.78204	0.50988	0.50832	150	150	B	B	biological fallback assumption	illumina	nextseq	unknown	poly_a	unknown	sc	single_cell_plate	smartseq	China	2024-04-08	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
35474	35474	SRR32818793	SRX28102226	SRS24458273	SRP572438	PRJNA1240805	Transcriptome profiling of tendon fibroblasts at the onset of embryonic muscle contraction reveals novel force responsive genes. FAC sorted tenocytes from 36 hpf vs. 48 hpf zebrafish embryos bulk RNAseq dataset	GSE292682	Transcriptome Analysis	Mechanical forces play a critical role in tendon development and function influencing cell behavior through mechanotransduction signaling pathways and subsequent extracellular matrix ECM remodeling. Here we investigate the molecular mechanisms by which tenocytes in developing zebrafish embryos respond to muscle contraction forces during the onset of swimming and cranial muscle activity. Using genome wide bulk RNA sequencing of FAC sorted tenocytes we identify novel tenocyte markers and genes involved in tendon mechanotransduction. Overall design: FAC sorted mCherry+ cells from 36 hpf and 48 hpf Tgscxa:mCherry zebrafish embryos labeling tendon fibroblasts tenocytes. 36 hpf embryos have not yet begun active muscle contraction and 48 hpf embryos are freely swimming indicating active muscle contraction onset.	pubmed:40145570	sorted mCherry+ tenocytes from 48 hpf Tgscxa:mCherry zebrafish embryos replicate 4	GSM8863184	source name:FAC sorted mCherry+ tenocytes from whole Tgscxa:mCherry embryos\|tissue:FAC sorted mCherry+ tenocytes from whole Tgscxa:mCherry embryos\|cell type:mCherry+ tenocytes\|genotype:WT Tgscxa:mCherry background\|treatment:48 hpf\|geo loc name:missing\|collection date:missing	sorted mCherry+ tenocytes from 48 hpf Tgscxa:mCherry zebrafish embryos replicate 4	Reads were mapped to zebrafish GRCz10 and quantified using STAR v2.5.2A and RSEM 1.2.31. Differential expression analysis were performed using DESeq2 v.1.30.1 Assembly: GRCz10 Supplementary files format and content: csv file includes DESeq2 normalized counts for all samples	FAC sorted mCherry+ tenocytes from whole Tgscxa:mCherry embryos		36 hpf and 48 hpf Tgscxa:mCherry zebrafish embryos labeling tendon fibroblasts/tenocytes were dissociated with Collagenase IV Gibco 17104019 at a concentration of 6.24 mg/ml without xxx addition at a temperature of 28C for roughly 40 min homogenizing every 5 min with a P1000 pipette. Cells were filtered through a 40um filter Pluriselect usa 43 10040 50. Dissociated cell suspensions were sorted on a Bio Rad FACS Aria II cell sorter. mCherry positive cells were gated and sorted for those expressing at high levels. RNEasy Micro Kit Qiagen 74004 was used for RNA extraction of cell lysates from FAC sorted cells and the Smart seq2 protocol was utilized for cDNA library construction. Libraries were sequenced using a Hi seq 4000 sequencer Illumina at a read depth of 35M reads per replicate.		tissue:FAC sorted mCherry+ tenocytes from whole Tgscxa:mCherry embryos\|cell type:mCherry+ tenocytes\|genotype:WT Tgscxa:mCherry background\|treatment:48 hpf	GSM8863184	GSM8863184: sorted mCherry+ tenocytes from 48 hpf Tgscxa:mCherry zebrafish embryos replicate 4; Danio rerio; RNA Seq	GSM8863184 r1	GSM8863184	1	36 hpf and 48 hpf Tgscxa:mCherry zebrafish embryos labeling tendon fibroblasts/tenocytes were dissociated with Collagenase IV Gibco 17104019 at a concentration of 6.24 mg/ml without xxx addition at a temperature of 28C for roughly 40 min homogenizing every 5 min with a P1000 pipette. Cells were filtered through a 40um filter Pluriselect usa 43 10040 50. Dissociated cell suspensions were sorted on a Bio Rad FACS Aria II cell sorter. mCherry positive cells were gated and sorted for those expressing at high levels. RNEasy Micro Kit Qiagen 74004 was used for RNA extraction of cell lysates from FAC sorted cells and the Smart seq2 protocol was utilized for cDNA library construction. Libraries were sequenced using a Hi seq 4000 sequencer Illumina at a read depth of 35M reads per replicate.		RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 4000	SRP572438		p48-4-READ1-Sequences.txt p48-4-READ2-Sequences.txt	fastq fastq	6624217400.0	33121087.0	GSM8863184 r1	0:100 1:100	A:1723006485;C:1604147349;G:1556891221;T:1739763662;N:408683	100	100	1723006485	1604147349	1556891221	1739763662	408683	SRX28102226	SRS24458273	SRA2098458	UC Irvine	UC Irvine												B	B	biological fallback assumption	illumina	hiseq_era	unknown	cdna_unspecified	unknown	sc	single_cell_plate	smartseq	United States	2025-03-23	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
35475	35475	SRR32818794	SRX28102225	SRS24458272	SRP572438	PRJNA1240805	Transcriptome profiling of tendon fibroblasts at the onset of embryonic muscle contraction reveals novel force responsive genes. FAC sorted tenocytes from 36 hpf vs. 48 hpf zebrafish embryos bulk RNAseq dataset	GSE292682	Transcriptome Analysis	Mechanical forces play a critical role in tendon development and function influencing cell behavior through mechanotransduction signaling pathways and subsequent extracellular matrix ECM remodeling. Here we investigate the molecular mechanisms by which tenocytes in developing zebrafish embryos respond to muscle contraction forces during the onset of swimming and cranial muscle activity. Using genome wide bulk RNA sequencing of FAC sorted tenocytes we identify novel tenocyte markers and genes involved in tendon mechanotransduction. Overall design: FAC sorted mCherry+ cells from 36 hpf and 48 hpf Tgscxa:mCherry zebrafish embryos labeling tendon fibroblasts tenocytes. 36 hpf embryos have not yet begun active muscle contraction and 48 hpf embryos are freely swimming indicating active muscle contraction onset.	pubmed:40145570	sorted mCherry+ tenocytes from 48 hpf Tgscxa:mCherry zebrafish embryos replicate 3	GSM8863183	source name:FAC sorted mCherry+ tenocytes from whole Tgscxa:mCherry embryos\|tissue:FAC sorted mCherry+ tenocytes from whole Tgscxa:mCherry embryos\|cell type:mCherry+ tenocytes\|genotype:WT Tgscxa:mCherry background\|treatment:48 hpf\|geo loc name:missing\|collection date:missing	sorted mCherry+ tenocytes from 48 hpf Tgscxa:mCherry zebrafish embryos replicate 3	Reads were mapped to zebrafish GRCz10 and quantified using STAR v2.5.2A and RSEM 1.2.31. Differential expression analysis were performed using DESeq2 v.1.30.1 Assembly: GRCz10 Supplementary files format and content: csv file includes DESeq2 normalized counts for all samples	FAC sorted mCherry+ tenocytes from whole Tgscxa:mCherry embryos		36 hpf and 48 hpf Tgscxa:mCherry zebrafish embryos labeling tendon fibroblasts/tenocytes were dissociated with Collagenase IV Gibco 17104019 at a concentration of 6.24 mg/ml without xxx addition at a temperature of 28C for roughly 40 min homogenizing every 5 min with a P1000 pipette. Cells were filtered through a 40um filter Pluriselect usa 43 10040 50. Dissociated cell suspensions were sorted on a Bio Rad FACS Aria II cell sorter. mCherry positive cells were gated and sorted for those expressing at high levels. RNEasy Micro Kit Qiagen 74004 was used for RNA extraction of cell lysates from FAC sorted cells and the Smart seq2 protocol was utilized for cDNA library construction. Libraries were sequenced using a Hi seq 4000 sequencer Illumina at a read depth of 35M reads per replicate.		tissue:FAC sorted mCherry+ tenocytes from whole Tgscxa:mCherry embryos\|cell type:mCherry+ tenocytes\|genotype:WT Tgscxa:mCherry background\|treatment:48 hpf	GSM8863183	GSM8863183: sorted mCherry+ tenocytes from 48 hpf Tgscxa:mCherry zebrafish embryos replicate 3; Danio rerio; RNA Seq	GSM8863183 r1	GSM8863183	1	36 hpf and 48 hpf Tgscxa:mCherry zebrafish embryos labeling tendon fibroblasts/tenocytes were dissociated with Collagenase IV Gibco 17104019 at a concentration of 6.24 mg/ml without xxx addition at a temperature of 28C for roughly 40 min homogenizing every 5 min with a P1000 pipette. Cells were filtered through a 40um filter Pluriselect usa 43 10040 50. Dissociated cell suspensions were sorted on a Bio Rad FACS Aria II cell sorter. mCherry positive cells were gated and sorted for those expressing at high levels. RNEasy Micro Kit Qiagen 74004 was used for RNA extraction of cell lysates from FAC sorted cells and the Smart seq2 protocol was utilized for cDNA library construction. Libraries were sequenced using a Hi seq 4000 sequencer Illumina at a read depth of 35M reads per replicate.		RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 4000	SRP572438		p48-3-READ1-Sequences.txt p48-3-READ2-Sequences.txt	fastq fastq	5127128400.0	25635642.0	GSM8863183 r1	0:100 1:100	A:1321647066;C:1250868808;G:1210591826;T:1343703952;N:316748	100	100	1321647066	1250868808	1210591826	1343703952	316748	SRX28102225	SRS24458272	SRA2098458	UC Irvine	UC Irvine												B	B	biological fallback assumption	illumina	hiseq_era	unknown	cdna_unspecified	unknown	sc	single_cell_plate	smartseq	United States	2025-03-23	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
35476	35476	SRR32818795	SRX28102224	SRS24458271	SRP572438	PRJNA1240805	Transcriptome profiling of tendon fibroblasts at the onset of embryonic muscle contraction reveals novel force responsive genes. FAC sorted tenocytes from 36 hpf vs. 48 hpf zebrafish embryos bulk RNAseq dataset	GSE292682	Transcriptome Analysis	Mechanical forces play a critical role in tendon development and function influencing cell behavior through mechanotransduction signaling pathways and subsequent extracellular matrix ECM remodeling. Here we investigate the molecular mechanisms by which tenocytes in developing zebrafish embryos respond to muscle contraction forces during the onset of swimming and cranial muscle activity. Using genome wide bulk RNA sequencing of FAC sorted tenocytes we identify novel tenocyte markers and genes involved in tendon mechanotransduction. Overall design: FAC sorted mCherry+ cells from 36 hpf and 48 hpf Tgscxa:mCherry zebrafish embryos labeling tendon fibroblasts tenocytes. 36 hpf embryos have not yet begun active muscle contraction and 48 hpf embryos are freely swimming indicating active muscle contraction onset.	pubmed:40145570	sorted mCherry+ tenocytes from 48 hpf Tgscxa:mCherry zebrafish embryos replicate 2	GSM8863182	source name:FAC sorted mCherry+ tenocytes from whole Tgscxa:mCherry embryos\|tissue:FAC sorted mCherry+ tenocytes from whole Tgscxa:mCherry embryos\|cell type:mCherry+ tenocytes\|genotype:WT Tgscxa:mCherry background\|treatment:48 hpf\|geo loc name:missing\|collection date:missing	sorted mCherry+ tenocytes from 48 hpf Tgscxa:mCherry zebrafish embryos replicate 2	Reads were mapped to zebrafish GRCz10 and quantified using STAR v2.5.2A and RSEM 1.2.31. Differential expression analysis were performed using DESeq2 v.1.30.1 Assembly: GRCz10 Supplementary files format and content: csv file includes DESeq2 normalized counts for all samples	FAC sorted mCherry+ tenocytes from whole Tgscxa:mCherry embryos		36 hpf and 48 hpf Tgscxa:mCherry zebrafish embryos labeling tendon fibroblasts/tenocytes were dissociated with Collagenase IV Gibco 17104019 at a concentration of 6.24 mg/ml without xxx addition at a temperature of 28C for roughly 40 min homogenizing every 5 min with a P1000 pipette. Cells were filtered through a 40um filter Pluriselect usa 43 10040 50. Dissociated cell suspensions were sorted on a Bio Rad FACS Aria II cell sorter. mCherry positive cells were gated and sorted for those expressing at high levels. RNEasy Micro Kit Qiagen 74004 was used for RNA extraction of cell lysates from FAC sorted cells and the Smart seq2 protocol was utilized for cDNA library construction. Libraries were sequenced using a Hi seq 4000 sequencer Illumina at a read depth of 35M reads per replicate.		tissue:FAC sorted mCherry+ tenocytes from whole Tgscxa:mCherry embryos\|cell type:mCherry+ tenocytes\|genotype:WT Tgscxa:mCherry background\|treatment:48 hpf	GSM8863182	GSM8863182: sorted mCherry+ tenocytes from 48 hpf Tgscxa:mCherry zebrafish embryos replicate 2; Danio rerio; RNA Seq	GSM8863182 r1	GSM8863182	1	36 hpf and 48 hpf Tgscxa:mCherry zebrafish embryos labeling tendon fibroblasts/tenocytes were dissociated with Collagenase IV Gibco 17104019 at a concentration of 6.24 mg/ml without xxx addition at a temperature of 28C for roughly 40 min homogenizing every 5 min with a P1000 pipette. Cells were filtered through a 40um filter Pluriselect usa 43 10040 50. Dissociated cell suspensions were sorted on a Bio Rad FACS Aria II cell sorter. mCherry positive cells were gated and sorted for those expressing at high levels. RNEasy Micro Kit Qiagen 74004 was used for RNA extraction of cell lysates from FAC sorted cells and the Smart seq2 protocol was utilized for cDNA library construction. Libraries were sequenced using a Hi seq 4000 sequencer Illumina at a read depth of 35M reads per replicate.		RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 4000	SRP572438		p48-2-READ1-Sequences.txt p48-2-READ2-Sequences.txt	fastq fastq	6832644200.0	34163221.0	GSM8863182 r1	0:100 1:100	A:1762283242;C:1668293067;G:1612470421;T:1789177247;N:420223	100	100	1762283242	1668293067	1612470421	1789177247	420223	SRX28102224	SRS24458271	SRA2098458	UC Irvine	UC Irvine												B	B	biological fallback assumption	illumina	hiseq_era	unknown	cdna_unspecified	unknown	sc	single_cell_plate	smartseq	United States	2025-03-23	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
35477	35477	SRR32818796	SRX28102223	SRS24458270	SRP572438	PRJNA1240805	Transcriptome profiling of tendon fibroblasts at the onset of embryonic muscle contraction reveals novel force responsive genes. FAC sorted tenocytes from 36 hpf vs. 48 hpf zebrafish embryos bulk RNAseq dataset	GSE292682	Transcriptome Analysis	Mechanical forces play a critical role in tendon development and function influencing cell behavior through mechanotransduction signaling pathways and subsequent extracellular matrix ECM remodeling. Here we investigate the molecular mechanisms by which tenocytes in developing zebrafish embryos respond to muscle contraction forces during the onset of swimming and cranial muscle activity. Using genome wide bulk RNA sequencing of FAC sorted tenocytes we identify novel tenocyte markers and genes involved in tendon mechanotransduction. Overall design: FAC sorted mCherry+ cells from 36 hpf and 48 hpf Tgscxa:mCherry zebrafish embryos labeling tendon fibroblasts tenocytes. 36 hpf embryos have not yet begun active muscle contraction and 48 hpf embryos are freely swimming indicating active muscle contraction onset.	pubmed:40145570	sorted mCherry+ tenocytes from 48 hpf Tgscxa:mCherry zebrafish embryos replicate 1	GSM8863181	source name:FAC sorted mCherry+ tenocytes from whole Tgscxa:mCherry embryos\|tissue:FAC sorted mCherry+ tenocytes from whole Tgscxa:mCherry embryos\|cell type:mCherry+ tenocytes\|genotype:WT Tgscxa:mCherry background\|treatment:48 hpf\|geo loc name:missing\|collection date:missing	sorted mCherry+ tenocytes from 48 hpf Tgscxa:mCherry zebrafish embryos replicate 1	Reads were mapped to zebrafish GRCz10 and quantified using STAR v2.5.2A and RSEM 1.2.31. Differential expression analysis were performed using DESeq2 v.1.30.1 Assembly: GRCz10 Supplementary files format and content: csv file includes DESeq2 normalized counts for all samples	FAC sorted mCherry+ tenocytes from whole Tgscxa:mCherry embryos		36 hpf and 48 hpf Tgscxa:mCherry zebrafish embryos labeling tendon fibroblasts/tenocytes were dissociated with Collagenase IV Gibco 17104019 at a concentration of 6.24 mg/ml without xxx addition at a temperature of 28C for roughly 40 min homogenizing every 5 min with a P1000 pipette. Cells were filtered through a 40um filter Pluriselect usa 43 10040 50. Dissociated cell suspensions were sorted on a Bio Rad FACS Aria II cell sorter. mCherry positive cells were gated and sorted for those expressing at high levels. RNEasy Micro Kit Qiagen 74004 was used for RNA extraction of cell lysates from FAC sorted cells and the Smart seq2 protocol was utilized for cDNA library construction. Libraries were sequenced using a Hi seq 4000 sequencer Illumina at a read depth of 35M reads per replicate.		tissue:FAC sorted mCherry+ tenocytes from whole Tgscxa:mCherry embryos\|cell type:mCherry+ tenocytes\|genotype:WT Tgscxa:mCherry background\|treatment:48 hpf	GSM8863181	GSM8863181: sorted mCherry+ tenocytes from 48 hpf Tgscxa:mCherry zebrafish embryos replicate 1; Danio rerio; RNA Seq	GSM8863181 r1	GSM8863181	1	36 hpf and 48 hpf Tgscxa:mCherry zebrafish embryos labeling tendon fibroblasts/tenocytes were dissociated with Collagenase IV Gibco 17104019 at a concentration of 6.24 mg/ml without xxx addition at a temperature of 28C for roughly 40 min homogenizing every 5 min with a P1000 pipette. Cells were filtered through a 40um filter Pluriselect usa 43 10040 50. Dissociated cell suspensions were sorted on a Bio Rad FACS Aria II cell sorter. mCherry positive cells were gated and sorted for those expressing at high levels. RNEasy Micro Kit Qiagen 74004 was used for RNA extraction of cell lysates from FAC sorted cells and the Smart seq2 protocol was utilized for cDNA library construction. Libraries were sequenced using a Hi seq 4000 sequencer Illumina at a read depth of 35M reads per replicate.		RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 4000	SRP572438		p48-1-READ1-Sequences.txt p48-1-READ2-Sequences.txt	fastq fastq	6861200200.0	34306001.0	GSM8863181 r1	0:100 1:100	A:1765390165;C:1682522416;G:1623224185;T:1789639564;N:423870	100	100	1765390165	1682522416	1623224185	1789639564	423870	SRX28102223	SRS24458270	SRA2098458	UC Irvine	UC Irvine												B	B	biological fallback assumption	illumina	hiseq_era	unknown	cdna_unspecified	unknown	sc	single_cell_plate	smartseq	United States	2025-03-23	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
35478	35478	SRR32818797	SRX28102222	SRS24458269	SRP572438	PRJNA1240805	Transcriptome profiling of tendon fibroblasts at the onset of embryonic muscle contraction reveals novel force responsive genes. FAC sorted tenocytes from 36 hpf vs. 48 hpf zebrafish embryos bulk RNAseq dataset	GSE292682	Transcriptome Analysis	Mechanical forces play a critical role in tendon development and function influencing cell behavior through mechanotransduction signaling pathways and subsequent extracellular matrix ECM remodeling. Here we investigate the molecular mechanisms by which tenocytes in developing zebrafish embryos respond to muscle contraction forces during the onset of swimming and cranial muscle activity. Using genome wide bulk RNA sequencing of FAC sorted tenocytes we identify novel tenocyte markers and genes involved in tendon mechanotransduction. Overall design: FAC sorted mCherry+ cells from 36 hpf and 48 hpf Tgscxa:mCherry zebrafish embryos labeling tendon fibroblasts tenocytes. 36 hpf embryos have not yet begun active muscle contraction and 48 hpf embryos are freely swimming indicating active muscle contraction onset.	pubmed:40145570	sorted mCherry+ tenocytes from 36 hpf Tgscxa:mCherry zebrafish embryos replicate 7	GSM8863180	source name:FAC sorted mCherry+ tenocytes from whole Tgscxa:mCherry embryos\|tissue:FAC sorted mCherry+ tenocytes from whole Tgscxa:mCherry embryos\|cell type:mCherry+ tenocytes\|genotype:WT Tgscxa:mCherry background\|treatment:36 hpf\|geo loc name:missing\|collection date:missing	sorted mCherry+ tenocytes from 36 hpf Tgscxa:mCherry zebrafish embryos replicate 7	Reads were mapped to zebrafish GRCz10 and quantified using STAR v2.5.2A and RSEM 1.2.31. Differential expression analysis were performed using DESeq2 v.1.30.1 Assembly: GRCz10 Supplementary files format and content: csv file includes DESeq2 normalized counts for all samples	FAC sorted mCherry+ tenocytes from whole Tgscxa:mCherry embryos		36 hpf and 48 hpf Tgscxa:mCherry zebrafish embryos labeling tendon fibroblasts/tenocytes were dissociated with Collagenase IV Gibco 17104019 at a concentration of 6.24 mg/ml without xxx addition at a temperature of 28C for roughly 40 min homogenizing every 5 min with a P1000 pipette. Cells were filtered through a 40um filter Pluriselect usa 43 10040 50. Dissociated cell suspensions were sorted on a Bio Rad FACS Aria II cell sorter. mCherry positive cells were gated and sorted for those expressing at high levels. RNEasy Micro Kit Qiagen 74004 was used for RNA extraction of cell lysates from FAC sorted cells and the Smart seq2 protocol was utilized for cDNA library construction. Libraries were sequenced using a Hi seq 4000 sequencer Illumina at a read depth of 35M reads per replicate.		tissue:FAC sorted mCherry+ tenocytes from whole Tgscxa:mCherry embryos\|cell type:mCherry+ tenocytes\|genotype:WT Tgscxa:mCherry background\|treatment:36 hpf	GSM8863180	GSM8863180: sorted mCherry+ tenocytes from 36 hpf Tgscxa:mCherry zebrafish embryos replicate 7; Danio rerio; RNA Seq	GSM8863180 r1	GSM8863180	1	36 hpf and 48 hpf Tgscxa:mCherry zebrafish embryos labeling tendon fibroblasts/tenocytes were dissociated with Collagenase IV Gibco 17104019 at a concentration of 6.24 mg/ml without xxx addition at a temperature of 28C for roughly 40 min homogenizing every 5 min with a P1000 pipette. Cells were filtered through a 40um filter Pluriselect usa 43 10040 50. Dissociated cell suspensions were sorted on a Bio Rad FACS Aria II cell sorter. mCherry positive cells were gated and sorted for those expressing at high levels. RNEasy Micro Kit Qiagen 74004 was used for RNA extraction of cell lysates from FAC sorted cells and the Smart seq2 protocol was utilized for cDNA library construction. Libraries were sequenced using a Hi seq 4000 sequencer Illumina at a read depth of 35M reads per replicate.		RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 4000	SRP572438		p36-7-READ1-Sequences.txt p36-7-READ2-Sequences.txt	fastq fastq	5620392000.0	28101960.0	GSM8863180 r1	0:100 1:100	A:1445709762;C:1377367138;G:1310017425;T:1486952834;N:344841	100	100	1445709762	1377367138	1310017425	1486952834	344841	SRX28102222	SRS24458269	SRA2098458	UC Irvine	UC Irvine												B	B	biological fallback assumption	illumina	hiseq_era	unknown	cdna_unspecified	unknown	sc	single_cell_plate	smartseq	United States	2025-03-23	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
35479	35479	SRR32818798	SRX28102221	SRS24458268	SRP572438	PRJNA1240805	Transcriptome profiling of tendon fibroblasts at the onset of embryonic muscle contraction reveals novel force responsive genes. FAC sorted tenocytes from 36 hpf vs. 48 hpf zebrafish embryos bulk RNAseq dataset	GSE292682	Transcriptome Analysis	Mechanical forces play a critical role in tendon development and function influencing cell behavior through mechanotransduction signaling pathways and subsequent extracellular matrix ECM remodeling. Here we investigate the molecular mechanisms by which tenocytes in developing zebrafish embryos respond to muscle contraction forces during the onset of swimming and cranial muscle activity. Using genome wide bulk RNA sequencing of FAC sorted tenocytes we identify novel tenocyte markers and genes involved in tendon mechanotransduction. Overall design: FAC sorted mCherry+ cells from 36 hpf and 48 hpf Tgscxa:mCherry zebrafish embryos labeling tendon fibroblasts tenocytes. 36 hpf embryos have not yet begun active muscle contraction and 48 hpf embryos are freely swimming indicating active muscle contraction onset.	pubmed:40145570	sorted mCherry+ tenocytes from 36 hpf Tgscxa:mCherry zebrafish embryos replicate 6	GSM8863179	source name:FAC sorted mCherry+ tenocytes from whole Tgscxa:mCherry embryos\|tissue:FAC sorted mCherry+ tenocytes from whole Tgscxa:mCherry embryos\|cell type:mCherry+ tenocytes\|genotype:WT Tgscxa:mCherry background\|treatment:36 hpf\|geo loc name:missing\|collection date:missing	sorted mCherry+ tenocytes from 36 hpf Tgscxa:mCherry zebrafish embryos replicate 6	Reads were mapped to zebrafish GRCz10 and quantified using STAR v2.5.2A and RSEM 1.2.31. Differential expression analysis were performed using DESeq2 v.1.30.1 Assembly: GRCz10 Supplementary files format and content: csv file includes DESeq2 normalized counts for all samples	FAC sorted mCherry+ tenocytes from whole Tgscxa:mCherry embryos		36 hpf and 48 hpf Tgscxa:mCherry zebrafish embryos labeling tendon fibroblasts/tenocytes were dissociated with Collagenase IV Gibco 17104019 at a concentration of 6.24 mg/ml without xxx addition at a temperature of 28C for roughly 40 min homogenizing every 5 min with a P1000 pipette. Cells were filtered through a 40um filter Pluriselect usa 43 10040 50. Dissociated cell suspensions were sorted on a Bio Rad FACS Aria II cell sorter. mCherry positive cells were gated and sorted for those expressing at high levels. RNEasy Micro Kit Qiagen 74004 was used for RNA extraction of cell lysates from FAC sorted cells and the Smart seq2 protocol was utilized for cDNA library construction. Libraries were sequenced using a Hi seq 4000 sequencer Illumina at a read depth of 35M reads per replicate.		tissue:FAC sorted mCherry+ tenocytes from whole Tgscxa:mCherry embryos\|cell type:mCherry+ tenocytes\|genotype:WT Tgscxa:mCherry background\|treatment:36 hpf	GSM8863179	GSM8863179: sorted mCherry+ tenocytes from 36 hpf Tgscxa:mCherry zebrafish embryos replicate 6; Danio rerio; RNA Seq	GSM8863179 r1	GSM8863179	1	36 hpf and 48 hpf Tgscxa:mCherry zebrafish embryos labeling tendon fibroblasts/tenocytes were dissociated with Collagenase IV Gibco 17104019 at a concentration of 6.24 mg/ml without xxx addition at a temperature of 28C for roughly 40 min homogenizing every 5 min with a P1000 pipette. Cells were filtered through a 40um filter Pluriselect usa 43 10040 50. Dissociated cell suspensions were sorted on a Bio Rad FACS Aria II cell sorter. mCherry positive cells were gated and sorted for those expressing at high levels. RNEasy Micro Kit Qiagen 74004 was used for RNA extraction of cell lysates from FAC sorted cells and the Smart seq2 protocol was utilized for cDNA library construction. Libraries were sequenced using a Hi seq 4000 sequencer Illumina at a read depth of 35M reads per replicate.		RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 4000	SRP572438		p36-6-READ1-Sequences.txt p36-6-READ2-Sequences.txt	fastq fastq	7568489000.0	37842445.0	GSM8863179 r1	0:100 1:100	A:1969395695;C:1827574975;G:1754405241;T:2016644374;N:468715	100	100	1969395695	1827574975	1754405241	2016644374	468715	SRX28102221	SRS24458268	SRA2098458	UC Irvine	UC Irvine												B	B	biological fallback assumption	illumina	hiseq_era	unknown	cdna_unspecified	unknown	sc	single_cell_plate	smartseq	United States	2025-03-23	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
35480	35480	SRR32818799	SRX28102220	SRS24458267	SRP572438	PRJNA1240805	Transcriptome profiling of tendon fibroblasts at the onset of embryonic muscle contraction reveals novel force responsive genes. FAC sorted tenocytes from 36 hpf vs. 48 hpf zebrafish embryos bulk RNAseq dataset	GSE292682	Transcriptome Analysis	Mechanical forces play a critical role in tendon development and function influencing cell behavior through mechanotransduction signaling pathways and subsequent extracellular matrix ECM remodeling. Here we investigate the molecular mechanisms by which tenocytes in developing zebrafish embryos respond to muscle contraction forces during the onset of swimming and cranial muscle activity. Using genome wide bulk RNA sequencing of FAC sorted tenocytes we identify novel tenocyte markers and genes involved in tendon mechanotransduction. Overall design: FAC sorted mCherry+ cells from 36 hpf and 48 hpf Tgscxa:mCherry zebrafish embryos labeling tendon fibroblasts tenocytes. 36 hpf embryos have not yet begun active muscle contraction and 48 hpf embryos are freely swimming indicating active muscle contraction onset.	pubmed:40145570	sorted mCherry+ tenocytes from 36 hpf Tgscxa:mCherry zebrafish embryos replicate 5	GSM8863178	source name:FAC sorted mCherry+ tenocytes from whole Tgscxa:mCherry embryos\|tissue:FAC sorted mCherry+ tenocytes from whole Tgscxa:mCherry embryos\|cell type:mCherry+ tenocytes\|genotype:WT Tgscxa:mCherry background\|treatment:36 hpf\|geo loc name:missing\|collection date:missing	sorted mCherry+ tenocytes from 36 hpf Tgscxa:mCherry zebrafish embryos replicate 5	Reads were mapped to zebrafish GRCz10 and quantified using STAR v2.5.2A and RSEM 1.2.31. Differential expression analysis were performed using DESeq2 v.1.30.1 Assembly: GRCz10 Supplementary files format and content: csv file includes DESeq2 normalized counts for all samples	FAC sorted mCherry+ tenocytes from whole Tgscxa:mCherry embryos		36 hpf and 48 hpf Tgscxa:mCherry zebrafish embryos labeling tendon fibroblasts/tenocytes were dissociated with Collagenase IV Gibco 17104019 at a concentration of 6.24 mg/ml without xxx addition at a temperature of 28C for roughly 40 min homogenizing every 5 min with a P1000 pipette. Cells were filtered through a 40um filter Pluriselect usa 43 10040 50. Dissociated cell suspensions were sorted on a Bio Rad FACS Aria II cell sorter. mCherry positive cells were gated and sorted for those expressing at high levels. RNEasy Micro Kit Qiagen 74004 was used for RNA extraction of cell lysates from FAC sorted cells and the Smart seq2 protocol was utilized for cDNA library construction. Libraries were sequenced using a Hi seq 4000 sequencer Illumina at a read depth of 35M reads per replicate.		tissue:FAC sorted mCherry+ tenocytes from whole Tgscxa:mCherry embryos\|cell type:mCherry+ tenocytes\|genotype:WT Tgscxa:mCherry background\|treatment:36 hpf	GSM8863178	GSM8863178: sorted mCherry+ tenocytes from 36 hpf Tgscxa:mCherry zebrafish embryos replicate 5; Danio rerio; RNA Seq	GSM8863178 r1	GSM8863178	1	36 hpf and 48 hpf Tgscxa:mCherry zebrafish embryos labeling tendon fibroblasts/tenocytes were dissociated with Collagenase IV Gibco 17104019 at a concentration of 6.24 mg/ml without xxx addition at a temperature of 28C for roughly 40 min homogenizing every 5 min with a P1000 pipette. Cells were filtered through a 40um filter Pluriselect usa 43 10040 50. Dissociated cell suspensions were sorted on a Bio Rad FACS Aria II cell sorter. mCherry positive cells were gated and sorted for those expressing at high levels. RNEasy Micro Kit Qiagen 74004 was used for RNA extraction of cell lysates from FAC sorted cells and the Smart seq2 protocol was utilized for cDNA library construction. Libraries were sequenced using a Hi seq 4000 sequencer Illumina at a read depth of 35M reads per replicate.		RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 4000	SRP572438		p36-5-READ1-Sequences.txt p36-5-READ2-Sequences.txt	fastq fastq	6657832800.0	33289164.0	GSM8863178 r1	0:100 1:100	A:1726677355;C:1613449895;G:1548951675;T:1768343661;N:410214	100	100	1726677355	1613449895	1548951675	1768343661	410214	SRX28102220	SRS24458267	SRA2098458	UC Irvine	UC Irvine												B	B	biological fallback assumption	illumina	hiseq_era	unknown	cdna_unspecified	unknown	sc	single_cell_plate	smartseq	United States	2025-03-23	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
35481	35481	SRR32818800	SRX28102219	SRS24458266	SRP572438	PRJNA1240805	Transcriptome profiling of tendon fibroblasts at the onset of embryonic muscle contraction reveals novel force responsive genes. FAC sorted tenocytes from 36 hpf vs. 48 hpf zebrafish embryos bulk RNAseq dataset	GSE292682	Transcriptome Analysis	Mechanical forces play a critical role in tendon development and function influencing cell behavior through mechanotransduction signaling pathways and subsequent extracellular matrix ECM remodeling. Here we investigate the molecular mechanisms by which tenocytes in developing zebrafish embryos respond to muscle contraction forces during the onset of swimming and cranial muscle activity. Using genome wide bulk RNA sequencing of FAC sorted tenocytes we identify novel tenocyte markers and genes involved in tendon mechanotransduction. Overall design: FAC sorted mCherry+ cells from 36 hpf and 48 hpf Tgscxa:mCherry zebrafish embryos labeling tendon fibroblasts tenocytes. 36 hpf embryos have not yet begun active muscle contraction and 48 hpf embryos are freely swimming indicating active muscle contraction onset.	pubmed:40145570	sorted mCherry+ tenocytes from 36 hpf Tgscxa:mCherry zebrafish embryos replicate 4	GSM8863177	source name:FAC sorted mCherry+ tenocytes from whole Tgscxa:mCherry embryos\|tissue:FAC sorted mCherry+ tenocytes from whole Tgscxa:mCherry embryos\|cell type:mCherry+ tenocytes\|genotype:WT Tgscxa:mCherry background\|treatment:36 hpf\|geo loc name:missing\|collection date:missing	sorted mCherry+ tenocytes from 36 hpf Tgscxa:mCherry zebrafish embryos replicate 4	Reads were mapped to zebrafish GRCz10 and quantified using STAR v2.5.2A and RSEM 1.2.31. Differential expression analysis were performed using DESeq2 v.1.30.1 Assembly: GRCz10 Supplementary files format and content: csv file includes DESeq2 normalized counts for all samples	FAC sorted mCherry+ tenocytes from whole Tgscxa:mCherry embryos		36 hpf and 48 hpf Tgscxa:mCherry zebrafish embryos labeling tendon fibroblasts/tenocytes were dissociated with Collagenase IV Gibco 17104019 at a concentration of 6.24 mg/ml without xxx addition at a temperature of 28C for roughly 40 min homogenizing every 5 min with a P1000 pipette. Cells were filtered through a 40um filter Pluriselect usa 43 10040 50. Dissociated cell suspensions were sorted on a Bio Rad FACS Aria II cell sorter. mCherry positive cells were gated and sorted for those expressing at high levels. RNEasy Micro Kit Qiagen 74004 was used for RNA extraction of cell lysates from FAC sorted cells and the Smart seq2 protocol was utilized for cDNA library construction. Libraries were sequenced using a Hi seq 4000 sequencer Illumina at a read depth of 35M reads per replicate.		tissue:FAC sorted mCherry+ tenocytes from whole Tgscxa:mCherry embryos\|cell type:mCherry+ tenocytes\|genotype:WT Tgscxa:mCherry background\|treatment:36 hpf	GSM8863177	GSM8863177: sorted mCherry+ tenocytes from 36 hpf Tgscxa:mCherry zebrafish embryos replicate 4; Danio rerio; RNA Seq	GSM8863177 r1	GSM8863177	1	36 hpf and 48 hpf Tgscxa:mCherry zebrafish embryos labeling tendon fibroblasts/tenocytes were dissociated with Collagenase IV Gibco 17104019 at a concentration of 6.24 mg/ml without xxx addition at a temperature of 28C for roughly 40 min homogenizing every 5 min with a P1000 pipette. Cells were filtered through a 40um filter Pluriselect usa 43 10040 50. Dissociated cell suspensions were sorted on a Bio Rad FACS Aria II cell sorter. mCherry positive cells were gated and sorted for those expressing at high levels. RNEasy Micro Kit Qiagen 74004 was used for RNA extraction of cell lysates from FAC sorted cells and the Smart seq2 protocol was utilized for cDNA library construction. Libraries were sequenced using a Hi seq 4000 sequencer Illumina at a read depth of 35M reads per replicate.		RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 4000	SRP572438		p36-4-READ1-Sequences.txt p36-4-READ2-Sequences.txt	fastq fastq	5153263800.0	25766319.0	GSM8863177 r1	0:100 1:100	A:1334474056;C:1251026482;G:1191417010;T:1376027996;N:318256	100	100	1334474056	1251026482	1191417010	1376027996	318256	SRX28102219	SRS24458266	SRA2098458	UC Irvine	UC Irvine												B	B	biological fallback assumption	illumina	hiseq_era	unknown	cdna_unspecified	unknown	sc	single_cell_plate	smartseq	United States	2025-03-23	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
35482	35482	SRR32818801	SRX28102218	SRS24458265	SRP572438	PRJNA1240805	Transcriptome profiling of tendon fibroblasts at the onset of embryonic muscle contraction reveals novel force responsive genes. FAC sorted tenocytes from 36 hpf vs. 48 hpf zebrafish embryos bulk RNAseq dataset	GSE292682	Transcriptome Analysis	Mechanical forces play a critical role in tendon development and function influencing cell behavior through mechanotransduction signaling pathways and subsequent extracellular matrix ECM remodeling. Here we investigate the molecular mechanisms by which tenocytes in developing zebrafish embryos respond to muscle contraction forces during the onset of swimming and cranial muscle activity. Using genome wide bulk RNA sequencing of FAC sorted tenocytes we identify novel tenocyte markers and genes involved in tendon mechanotransduction. Overall design: FAC sorted mCherry+ cells from 36 hpf and 48 hpf Tgscxa:mCherry zebrafish embryos labeling tendon fibroblasts tenocytes. 36 hpf embryos have not yet begun active muscle contraction and 48 hpf embryos are freely swimming indicating active muscle contraction onset.	pubmed:40145570	sorted mCherry+ tenocytes from 36 hpf Tgscxa:mCherry zebrafish embryos replicate 3	GSM8863176	source name:FAC sorted mCherry+ tenocytes from whole Tgscxa:mCherry embryos\|tissue:FAC sorted mCherry+ tenocytes from whole Tgscxa:mCherry embryos\|cell type:mCherry+ tenocytes\|genotype:WT Tgscxa:mCherry background\|treatment:36 hpf\|geo loc name:missing\|collection date:missing	sorted mCherry+ tenocytes from 36 hpf Tgscxa:mCherry zebrafish embryos replicate 3	Reads were mapped to zebrafish GRCz10 and quantified using STAR v2.5.2A and RSEM 1.2.31. Differential expression analysis were performed using DESeq2 v.1.30.1 Assembly: GRCz10 Supplementary files format and content: csv file includes DESeq2 normalized counts for all samples	FAC sorted mCherry+ tenocytes from whole Tgscxa:mCherry embryos		36 hpf and 48 hpf Tgscxa:mCherry zebrafish embryos labeling tendon fibroblasts/tenocytes were dissociated with Collagenase IV Gibco 17104019 at a concentration of 6.24 mg/ml without xxx addition at a temperature of 28C for roughly 40 min homogenizing every 5 min with a P1000 pipette. Cells were filtered through a 40um filter Pluriselect usa 43 10040 50. Dissociated cell suspensions were sorted on a Bio Rad FACS Aria II cell sorter. mCherry positive cells were gated and sorted for those expressing at high levels. RNEasy Micro Kit Qiagen 74004 was used for RNA extraction of cell lysates from FAC sorted cells and the Smart seq2 protocol was utilized for cDNA library construction. Libraries were sequenced using a Hi seq 4000 sequencer Illumina at a read depth of 35M reads per replicate.		tissue:FAC sorted mCherry+ tenocytes from whole Tgscxa:mCherry embryos\|cell type:mCherry+ tenocytes\|genotype:WT Tgscxa:mCherry background\|treatment:36 hpf	GSM8863176	GSM8863176: sorted mCherry+ tenocytes from 36 hpf Tgscxa:mCherry zebrafish embryos replicate 3; Danio rerio; RNA Seq	GSM8863176 r1	GSM8863176	1	36 hpf and 48 hpf Tgscxa:mCherry zebrafish embryos labeling tendon fibroblasts/tenocytes were dissociated with Collagenase IV Gibco 17104019 at a concentration of 6.24 mg/ml without xxx addition at a temperature of 28C for roughly 40 min homogenizing every 5 min with a P1000 pipette. Cells were filtered through a 40um filter Pluriselect usa 43 10040 50. Dissociated cell suspensions were sorted on a Bio Rad FACS Aria II cell sorter. mCherry positive cells were gated and sorted for those expressing at high levels. RNEasy Micro Kit Qiagen 74004 was used for RNA extraction of cell lysates from FAC sorted cells and the Smart seq2 protocol was utilized for cDNA library construction. Libraries were sequenced using a Hi seq 4000 sequencer Illumina at a read depth of 35M reads per replicate.		RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 4000	SRP572438		p36-3-READ1-Sequences.txt p36-3-READ2-Sequences.txt	fastq fastq	6089823000.0	30449115.0	GSM8863176 r1	0:100 1:100	A:1565216204;C:1493655703;G:1438012972;T:1592561753;N:376368	100	100	1565216204	1493655703	1438012972	1592561753	376368	SRX28102218	SRS24458265	SRA2098458	UC Irvine	UC Irvine												B	B	biological fallback assumption	illumina	hiseq_era	unknown	cdna_unspecified	unknown	sc	single_cell_plate	smartseq	United States	2025-03-23	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
35483	35483	SRR32818802	SRX28102217	SRS24458264	SRP572438	PRJNA1240805	Transcriptome profiling of tendon fibroblasts at the onset of embryonic muscle contraction reveals novel force responsive genes. FAC sorted tenocytes from 36 hpf vs. 48 hpf zebrafish embryos bulk RNAseq dataset	GSE292682	Transcriptome Analysis	Mechanical forces play a critical role in tendon development and function influencing cell behavior through mechanotransduction signaling pathways and subsequent extracellular matrix ECM remodeling. Here we investigate the molecular mechanisms by which tenocytes in developing zebrafish embryos respond to muscle contraction forces during the onset of swimming and cranial muscle activity. Using genome wide bulk RNA sequencing of FAC sorted tenocytes we identify novel tenocyte markers and genes involved in tendon mechanotransduction. Overall design: FAC sorted mCherry+ cells from 36 hpf and 48 hpf Tgscxa:mCherry zebrafish embryos labeling tendon fibroblasts tenocytes. 36 hpf embryos have not yet begun active muscle contraction and 48 hpf embryos are freely swimming indicating active muscle contraction onset.	pubmed:40145570	sorted mCherry+ tenocytes from 36 hpf Tgscxa:mCherry zebrafish embryos replicate 2	GSM8863175	source name:FAC sorted mCherry+ tenocytes from whole Tgscxa:mCherry embryos\|tissue:FAC sorted mCherry+ tenocytes from whole Tgscxa:mCherry embryos\|cell type:mCherry+ tenocytes\|genotype:WT Tgscxa:mCherry background\|treatment:36 hpf\|geo loc name:missing\|collection date:missing	sorted mCherry+ tenocytes from 36 hpf Tgscxa:mCherry zebrafish embryos replicate 2	Reads were mapped to zebrafish GRCz10 and quantified using STAR v2.5.2A and RSEM 1.2.31. Differential expression analysis were performed using DESeq2 v.1.30.1 Assembly: GRCz10 Supplementary files format and content: csv file includes DESeq2 normalized counts for all samples	FAC sorted mCherry+ tenocytes from whole Tgscxa:mCherry embryos		36 hpf and 48 hpf Tgscxa:mCherry zebrafish embryos labeling tendon fibroblasts/tenocytes were dissociated with Collagenase IV Gibco 17104019 at a concentration of 6.24 mg/ml without xxx addition at a temperature of 28C for roughly 40 min homogenizing every 5 min with a P1000 pipette. Cells were filtered through a 40um filter Pluriselect usa 43 10040 50. Dissociated cell suspensions were sorted on a Bio Rad FACS Aria II cell sorter. mCherry positive cells were gated and sorted for those expressing at high levels. RNEasy Micro Kit Qiagen 74004 was used for RNA extraction of cell lysates from FAC sorted cells and the Smart seq2 protocol was utilized for cDNA library construction. Libraries were sequenced using a Hi seq 4000 sequencer Illumina at a read depth of 35M reads per replicate.		tissue:FAC sorted mCherry+ tenocytes from whole Tgscxa:mCherry embryos\|cell type:mCherry+ tenocytes\|genotype:WT Tgscxa:mCherry background\|treatment:36 hpf	GSM8863175	GSM8863175: sorted mCherry+ tenocytes from 36 hpf Tgscxa:mCherry zebrafish embryos replicate 2; Danio rerio; RNA Seq	GSM8863175 r1	GSM8863175	1	36 hpf and 48 hpf Tgscxa:mCherry zebrafish embryos labeling tendon fibroblasts/tenocytes were dissociated with Collagenase IV Gibco 17104019 at a concentration of 6.24 mg/ml without xxx addition at a temperature of 28C for roughly 40 min homogenizing every 5 min with a P1000 pipette. Cells were filtered through a 40um filter Pluriselect usa 43 10040 50. Dissociated cell suspensions were sorted on a Bio Rad FACS Aria II cell sorter. mCherry positive cells were gated and sorted for those expressing at high levels. RNEasy Micro Kit Qiagen 74004 was used for RNA extraction of cell lysates from FAC sorted cells and the Smart seq2 protocol was utilized for cDNA library construction. Libraries were sequenced using a Hi seq 4000 sequencer Illumina at a read depth of 35M reads per replicate.		RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 4000	SRP572438		p36-2-READ1-Sequences.txt p36-2-READ2-Sequences.txt	fastq fastq	5575798600.0	27878993.0	GSM8863175 r1	0:100 1:100	A:1443367045;C:1355595610;G:1298738880;T:1477753374;N:343691	100	100	1443367045	1355595610	1298738880	1477753374	343691	SRX28102217	SRS24458264	SRA2098458	UC Irvine	UC Irvine												B	B	biological fallback assumption	illumina	hiseq_era	unknown	cdna_unspecified	unknown	sc	single_cell_plate	smartseq	United States	2025-03-23	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
35484	35484	SRR32818803	SRX28102216	SRS24458263	SRP572438	PRJNA1240805	Transcriptome profiling of tendon fibroblasts at the onset of embryonic muscle contraction reveals novel force responsive genes. FAC sorted tenocytes from 36 hpf vs. 48 hpf zebrafish embryos bulk RNAseq dataset	GSE292682	Transcriptome Analysis	Mechanical forces play a critical role in tendon development and function influencing cell behavior through mechanotransduction signaling pathways and subsequent extracellular matrix ECM remodeling. Here we investigate the molecular mechanisms by which tenocytes in developing zebrafish embryos respond to muscle contraction forces during the onset of swimming and cranial muscle activity. Using genome wide bulk RNA sequencing of FAC sorted tenocytes we identify novel tenocyte markers and genes involved in tendon mechanotransduction. Overall design: FAC sorted mCherry+ cells from 36 hpf and 48 hpf Tgscxa:mCherry zebrafish embryos labeling tendon fibroblasts tenocytes. 36 hpf embryos have not yet begun active muscle contraction and 48 hpf embryos are freely swimming indicating active muscle contraction onset.	pubmed:40145570	sorted mCherry+ tenocytes from 36 hpf Tgscxa:mCherry zebrafish embryos replicate 1	GSM8863174	source name:FAC sorted mCherry+ tenocytes from whole Tgscxa:mCherry embryos\|tissue:FAC sorted mCherry+ tenocytes from whole Tgscxa:mCherry embryos\|cell type:mCherry+ tenocytes\|genotype:WT Tgscxa:mCherry background\|treatment:36 hpf\|geo loc name:missing\|collection date:missing	sorted mCherry+ tenocytes from 36 hpf Tgscxa:mCherry zebrafish embryos replicate 1	Reads were mapped to zebrafish GRCz10 and quantified using STAR v2.5.2A and RSEM 1.2.31. Differential expression analysis were performed using DESeq2 v.1.30.1 Assembly: GRCz10 Supplementary files format and content: csv file includes DESeq2 normalized counts for all samples	FAC sorted mCherry+ tenocytes from whole Tgscxa:mCherry embryos		36 hpf and 48 hpf Tgscxa:mCherry zebrafish embryos labeling tendon fibroblasts/tenocytes were dissociated with Collagenase IV Gibco 17104019 at a concentration of 6.24 mg/ml without xxx addition at a temperature of 28C for roughly 40 min homogenizing every 5 min with a P1000 pipette. Cells were filtered through a 40um filter Pluriselect usa 43 10040 50. Dissociated cell suspensions were sorted on a Bio Rad FACS Aria II cell sorter. mCherry positive cells were gated and sorted for those expressing at high levels. RNEasy Micro Kit Qiagen 74004 was used for RNA extraction of cell lysates from FAC sorted cells and the Smart seq2 protocol was utilized for cDNA library construction. Libraries were sequenced using a Hi seq 4000 sequencer Illumina at a read depth of 35M reads per replicate.		tissue:FAC sorted mCherry+ tenocytes from whole Tgscxa:mCherry embryos\|cell type:mCherry+ tenocytes\|genotype:WT Tgscxa:mCherry background\|treatment:36 hpf	GSM8863174	GSM8863174: sorted mCherry+ tenocytes from 36 hpf Tgscxa:mCherry zebrafish embryos replicate 1; Danio rerio; RNA Seq	GSM8863174 r1	GSM8863174	1	36 hpf and 48 hpf Tgscxa:mCherry zebrafish embryos labeling tendon fibroblasts/tenocytes were dissociated with Collagenase IV Gibco 17104019 at a concentration of 6.24 mg/ml without xxx addition at a temperature of 28C for roughly 40 min homogenizing every 5 min with a P1000 pipette. Cells were filtered through a 40um filter Pluriselect usa 43 10040 50. Dissociated cell suspensions were sorted on a Bio Rad FACS Aria II cell sorter. mCherry positive cells were gated and sorted for those expressing at high levels. RNEasy Micro Kit Qiagen 74004 was used for RNA extraction of cell lysates from FAC sorted cells and the Smart seq2 protocol was utilized for cDNA library construction. Libraries were sequenced using a Hi seq 4000 sequencer Illumina at a read depth of 35M reads per replicate.		RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 4000	SRP572438		p36-1-READ1-Sequences.txt p36-1-READ2-Sequences.txt	fastq fastq	6686065200.0	33430326.0	GSM8863174 r1	0:100 1:100	A:1705825267;C:1656952667;G:1564381164;T:1758493189;N:412913	100	100	1705825267	1656952667	1564381164	1758493189	412913	SRX28102216	SRS24458263	SRA2098458	UC Irvine	UC Irvine												B	B	biological fallback assumption	illumina	hiseq_era	unknown	cdna_unspecified	unknown	sc	single_cell_plate	smartseq	United States	2025-03-23	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38457	38457	SRR1873001	SRX914893	SRS869409	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S366	GSM1629126	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S366	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629126	GSM1629126: zf3 dis S366; Danio rerio; RNA Seq	GSM1629126		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629126	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S366.R2.fq zf3_dis_S366.R1.fq	fastq fastq	13953150.0	279063.0	GSM1629126 r1	0:25 1:25	A:3638478;C:3221565;G:3250022;T:3754586;N:88499	25	25	3638478	3221565	3250022	3754586	88499	SRX914893	SRS869409	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.68547	0.72347	0.17435	0.1776	0.89883	0.87645	0.78642	0.70697	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38458	38458	SRR1873000	SRX914892	SRS869415	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S365	GSM1629125	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S365	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629125	GSM1629125: zf3 dis S365; Danio rerio; RNA Seq	GSM1629125		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629125	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S365.R1.fq zf3_dis_S365.R2.fq	fastq fastq	18319750.0	366395.0	GSM1629125 r1	0:25 1:25	A:4713431;C:4306960;G:4405843;T:4784636;N:108880	25	25	4713431	4306960	4405843	4784636	108880	SRX914892	SRS869415	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.45031	0.41097	0.17574	0.16675	0.91098	0.92383	0.57619	0.61898	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38459	38459	SRR1872999	SRX914891	SRS869411	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S364	GSM1629124	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S364	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629124	GSM1629124: zf3 dis S364; Danio rerio; RNA Seq	GSM1629124		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629124	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S364.R1.fq zf3_dis_S364.R2.fq	fastq fastq	16291400.0	325828.0	GSM1629124 r1	0:25 1:25	A:4127905;C:3884549;G:3934831;T:4281925;N:62190	25	25	4127905	3884549	3934831	4281925	62190	SRX914891	SRS869411	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.64233	0.58456	0.18268	0.16828	0.89079	0.91056	0.69046	0.76873	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38460	38460	SRR1872998	SRX914890	SRS869413	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S363	GSM1629123	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S363	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629123	GSM1629123: zf3 dis S363; Danio rerio; RNA Seq	GSM1629123		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629123	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S363.R1.fq zf3_dis_S363.R2.fq	fastq fastq	24979500.0	499590.0	GSM1629123 r1	0:25 1:25	A:6522125;C:5838067;G:5906327;T:6630372;N:82609	25	25	6522125	5838067	5906327	6630372	82609	SRX914890	SRS869413	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.78356	0.7193	0.35183	0.32737	0.85096	0.86953	0.65229	0.72787	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38461	38461	SRR1872997	SRX914889	SRS869410	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S362	GSM1629122	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S362	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629122	GSM1629122: zf3 dis S362; Danio rerio; RNA Seq	GSM1629122		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629122	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S362.R1.fq zf3_dis_S362.R2.fq	fastq fastq	25961150.0	519223.0	GSM1629122 r1	0:25 1:25	A:6513860;C:6274276;G:6330224;T:6742925;N:99865	25	25	6513860	6274276	6330224	6742925	99865	SRX914889	SRS869410	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.64392	0.70212	0.20822	0.2179	0.90449	0.88544	0.7447	0.66953	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38462	38462	SRR1872996	SRX914888	SRS869414	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S361	GSM1629121	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S361	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629121	GSM1629121: zf3 dis S361; Danio rerio; RNA Seq	GSM1629121		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629121	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S361.R1.fq zf3_dis_S361.R2.fq	fastq fastq	31158400.0	623168.0	GSM1629121 r1	0:25 1:25	A:8003929;C:7345167;G:7419409;T:8283464;N:106431	25	25	8003929	7345167	7419409	8283464	106431	SRX914888	SRS869414	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.69021	0.64535	0.18688	0.18054	0.86624	0.88789	0.69784	0.78517	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38463	38463	SRR1872995	SRX914887	SRS869416	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S360	GSM1629120	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S360	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629120	GSM1629120: zf3 dis S360; Danio rerio; RNA Seq	GSM1629120		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629120	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S360.R1.fq zf3_dis_S360.R2.fq	fastq fastq	65650.0	1313.0	GSM1629120 r1	0:25 1:25	A:16756;C:15635;G:13912;T:19083;N:264	25	25	16756	15635	13912	19083	264	SRX914887	SRS869416	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.6792	0.7566	0.21098	0.22718	0.99823	0.99667	0.76729	0.68093	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38464	38464	SRR1872994	SRX914886	SRS869408	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S359	GSM1629119	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S359	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629119	GSM1629119: zf3 dis S359; Danio rerio; RNA Seq	GSM1629119		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629119	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S359.R1.fq zf3_dis_S359.R2.fq	fastq fastq	11699050.0	233981.0	GSM1629119 r1	0:25 1:25	A:3029078;C:2720012;G:2748834;T:3147149;N:53977	25	25	3029078	2720012	2748834	3147149	53977	SRX914886	SRS869408	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.73706	0.69089	0.16704	0.16123	0.88795	0.90696	0.71161	0.78714	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38465	38465	SRR1872993	SRX914885	SRS869417	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S358	GSM1629118	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S358	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629118	GSM1629118: zf3 dis S358; Danio rerio; RNA Seq	GSM1629118		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629118	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S358.R2.fq zf3_dis_S358.R1.fq	fastq fastq	9555800.0	191116.0	GSM1629118 r1	0:25 1:25	A:2455617;C:2236174;G:2275012;T:2547459;N:41538	25	25	2455617	2236174	2275012	2547459	41538	SRX914885	SRS869417	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.71513	0.67049	0.20389	0.18909	0.90289	0.91758	0.72281	0.78587	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38466	38466	SRR1872992	SRX914884	SRS869420	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S357	GSM1629117	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S357	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629117	GSM1629117: zf3 dis S357; Danio rerio; RNA Seq	GSM1629117		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629117	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S357.R1.fq zf3_dis_S357.R2.fq	fastq fastq	6685400.0	133708.0	GSM1629117 r1	0:25 1:25	A:1794401;C:1506998;G:1499676;T:1830754;N:53571	25	25	1794401	1506998	1499676	1830754	53571	SRX914884	SRS869420	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.51579	0.51107	0.13925	0.12582	0.95089	0.95828	0.33071	0.8304	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38467	38467	SRR1872991	SRX914883	SRS869418	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S356	GSM1629116	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S356	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629116	GSM1629116: zf3 dis S356; Danio rerio; RNA Seq	GSM1629116		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629116	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S356.R2.fq zf3_dis_S356.R1.fq	fastq fastq	11637750.0	232755.0	GSM1629116 r1	0:25 1:25	A:2892151;C:2796777;G:2856920;T:3041989;N:49913	25	25	2892151	2796777	2856920	3041989	49913	SRX914883	SRS869418	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.78167	0.71632	0.23745	0.21955	0.89262	0.91054	0.67329	0.74109	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38468	38468	SRR1872990	SRX914882	SRS869419	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S355	GSM1629115	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S355	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629115	GSM1629115: zf3 dis S355; Danio rerio; RNA Seq	GSM1629115		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629115	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S355.R2.fq zf3_dis_S355.R1.fq	fastq fastq	16269550.0	325391.0	GSM1629115 r1	0:25 1:25	A:4085539;C:3887330;G:3971738;T:4261948;N:62995	25	25	4085539	3887330	3971738	4261948	62995	SRX914882	SRS869419	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.67445	0.74201	0.22487	0.24114	0.90473	0.88637	0.73772	0.66677	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38469	38469	SRR1872989	SRX914881	SRS869421	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S354	GSM1629114	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S354	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629114	GSM1629114: zf3 dis S354; Danio rerio; RNA Seq	GSM1629114		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629114	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S354.R1.fq zf3_dis_S354.R2.fq	fastq fastq	17585450.0	351709.0	GSM1629114 r1	0:25 1:25	A:4659589;C:4039300;G:4045812;T:4733042;N:107707	25	25	4659589	4039300	4045812	4733042	107707	SRX914881	SRS869421	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.4878	0.45754	0.15743	0.15052	0.91849	0.9321	0.70142	0.77697	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38470	38470	SRR1872988	SRX914880	SRS869422	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S353	GSM1629113	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S353	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629113	GSM1629113: zf3 dis S353; Danio rerio; RNA Seq	GSM1629113		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629113	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S353.R2.fq zf3_dis_S353.R1.fq	fastq fastq	14485550.0	289711.0	GSM1629113 r1	0:25 1:25	A:3738800;C:3399635;G:3435347;T:3844352;N:67416	25	25	3738800	3399635	3435347	3844352	67416	SRX914880	SRS869422	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.62055	0.56082	0.17925	0.17361	0.88698	0.90761	0.63705	0.72713	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38471	38471	SRR1872987	SRX914879	SRS869427	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S352	GSM1629112	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S352	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629112	GSM1629112: zf3 dis S352; Danio rerio; RNA Seq	GSM1629112		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629112	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S352.R1.fq zf3_dis_S352.R2.fq	fastq fastq	20656250.0	413125.0	GSM1629112 r1	0:25 1:25	A:5409278;C:4778324;G:4823018;T:5565607;N:80023	25	25	5409278	4778324	4823018	5565607	80023	SRX914879	SRS869427	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.7596	0.69086	0.29648	0.27453	0.84843	0.87265	0.65665	0.73501	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38472	38472	SRR1872986	SRX914878	SRS869407	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S351	GSM1629111	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S351	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629111	GSM1629111: zf3 dis S351; Danio rerio; RNA Seq	GSM1629111		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629111	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S351.R2.fq zf3_dis_S351.R1.fq	fastq fastq	18878500.0	377570.0	GSM1629111 r1	0:25 1:25	A:4794224;C:4526355;G:4554783;T:4925799;N:77339	25	25	4794224	4526355	4554783	4925799	77339	SRX914878	SRS869407	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.59735	0.65152	0.17597	0.18275	0.91204	0.89367	0.73252	0.66132	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38473	38473	SRR1872985	SRX914877	SRS869426	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S350	GSM1629110	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S350	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629110	GSM1629110: zf3 dis S350; Danio rerio; RNA Seq	GSM1629110		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629110	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S350.R2.fq zf3_dis_S350.R1.fq	fastq fastq	18015150.0	360303.0	GSM1629110 r1	0:25 1:25	A:4757233;C:4120910;G:4152471;T:4865140;N:119396	25	25	4757233	4120910	4152471	4865140	119396	SRX914877	SRS869426	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.59623	0.55628	0.16688	0.15933	0.89016	0.90946	0.68088	0.39729	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38474	38474	SRR1872984	SRX914876	SRS869425	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S349	GSM1629109	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S349	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629109	GSM1629109: zf3 dis S349; Danio rerio; RNA Seq	GSM1629109		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629109	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S349.R1.fq zf3_dis_S349.R2.fq	fastq fastq	18552750.0	371055.0	GSM1629109 r1	0:25 1:25	A:4713675;C:4414123;G:4487632;T:4865002;N:72318	25	25	4713675	4414123	4487632	4865002	72318	SRX914876	SRS869425	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.73956	0.66864	0.29185	0.26915	0.8728	0.89185	0.64518	0.71455	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38475	38475	SRR1872983	SRX914875	SRS869424	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S348	GSM1629108	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S348	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629108	GSM1629108: zf3 dis S348; Danio rerio; RNA Seq	GSM1629108		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629108	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S348.R1.fq zf3_dis_S348.R2.fq	fastq fastq	18364000.0	367280.0	GSM1629108 r1	0:25 1:25	A:4467688;C:4489708;G:4651022;T:4688256;N:67326	25	25	4467688	4489708	4651022	4688256	67326	SRX914875	SRS869424	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.82371	0.74488	0.27118	0.25268	0.86758	0.8886	0.65639	0.72281	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38476	38476	SRR1872982	SRX914874	SRS869429	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S347	GSM1629107	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S347	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629107	GSM1629107: zf3 dis S347; Danio rerio; RNA Seq	GSM1629107		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629107	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S347.R1.fq zf3_dis_S347.R2.fq	fastq fastq	21064200.0	421284.0	GSM1629107 r1	0:25 1:25	A:5412410;C:4977004;G:4997675;T:5572725;N:104386	25	25	5412410	4977004	4997675	5572725	104386	SRX914874	SRS869429	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.5825	0.54487	0.15602	0.14912	0.89053	0.91094	0.71218	0.79095	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38477	38477	SRR1872981	SRX914873	SRS869428	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S346	GSM1629106	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S346	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629106	GSM1629106: zf3 dis S346; Danio rerio; RNA Seq	GSM1629106		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629106	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S346.R1.fq zf3_dis_S346.R2.fq	fastq fastq	14579650.0	291593.0	GSM1629106 r1	0:25 1:25	A:3652547;C:3477071;G:3552653;T:3844470;N:52909	25	25	3652547	3477071	3552653	3844470	52909	SRX914873	SRS869428	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.84569	0.77715	0.24465	0.23283	0.87651	0.89706	0.69586	0.77338	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38478	38478	SRR1872980	SRX914872	SRS869430	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S345	GSM1629105	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S345	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629105	GSM1629105: zf3 dis S345; Danio rerio; RNA Seq	GSM1629105		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629105	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S345.R1.fq zf3_dis_S345.R2.fq	fastq fastq	19894100.0	397882.0	GSM1629105 r1	0:25 1:25	A:4978487;C:4785641;G:4877454;T:5178846;N:73672	25	25	4978487	4785641	4877454	5178846	73672	SRX914872	SRS869430	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.80074	0.73644	0.28821	0.27089	0.87452	0.89213	0.68194	0.75059	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38479	38479	SRR1872979	SRX914871	SRS869433	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S344	GSM1629104	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S344	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629104	GSM1629104: zf3 dis S344; Danio rerio; RNA Seq	GSM1629104		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629104	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S344.R2.fq zf3_dis_S344.R1.fq	fastq fastq	21335200.0	426704.0	GSM1629104 r1	0:25 1:25	A:5277911;C:5144153;G:5264908;T:5565367;N:82861	25	25	5277911	5144153	5264908	5565367	82861	SRX914871	SRS869433	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.72171	0.80382	0.21052	0.21995	0.89061	0.86695	0.72932	0.64495	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38480	38480	SRR1872978	SRX914870	SRS869431	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S343	GSM1629103	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S343	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629103	GSM1629103: zf3 dis S343; Danio rerio; RNA Seq	GSM1629103		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629103	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S343.R1.fq zf3_dis_S343.R2.fq	fastq fastq	19911000.0	398220.0	GSM1629103 r1	0:25 1:25	A:4923353;C:4815086;G:4933744;T:5166820;N:71997	25	25	4923353	4815086	4933744	5166820	71997	SRX914870	SRS869431	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.79232	0.71735	0.23984	0.22867	0.88943	0.91074	0.64925	0.73421	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38481	38481	SRR1872977	SRX914869	SRS869432	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S1800	GSM1629102	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S1800	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629102	GSM1629102: zf3 dis S1800; Danio rerio; RNA Seq	GSM1629102		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629102	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S1800.R1.fq zf3_dis_S1800.R2.fq	fastq fastq	10790700.0	215814.0	GSM1629102 r1	0:25 1:25	A:2721576;C:2556263;G:2606994;T:2865815;N:40052	25	25	2721576	2556263	2606994	2865815	40052	SRX914869	SRS869432	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.76882	0.8569	0.22815	0.24284	0.90851	0.88925	0.73921	0.67204	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38482	38482	SRR1872976	SRX914868	SRS869434	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S1799	GSM1629101	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S1799	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629101	GSM1629101: zf3 dis S1799; Danio rerio; RNA Seq	GSM1629101		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629101	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S1799.R1.fq zf3_dis_S1799.R2.fq	fastq fastq	5593150.0	111863.0	GSM1629101 r1	0:25 1:25	A:1395285;C:1345325;G:1376091;T:1455988;N:20461	25	25	1395285	1345325	1376091	1455988	20461	SRX914868	SRS869434	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.85141	0.7543	0.26701	0.24608	0.91443	0.93028	0.65519	0.7383	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38483	38483	SRR1872975	SRX914867	SRS869435	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S1798	GSM1629100	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S1798	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629100	GSM1629100: zf3 dis S1798; Danio rerio; RNA Seq	GSM1629100		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629100	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S1798.R2.fq zf3_dis_S1798.R1.fq	fastq fastq	10414350.0	208287.0	GSM1629100 r1	0:25 1:25	A:2580685;C:2509293;G:2567981;T:2715865;N:40526	25	25	2580685	2509293	2567981	2715865	40526	SRX914867	SRS869435	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.86357	0.77003	0.26639	0.24899	0.88876	0.90916	0.65432	0.73207	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38484	38484	SRR1872974	SRX914866	SRS869436	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S1797	GSM1629099	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S1797	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629099	GSM1629099: zf3 dis S1797; Danio rerio; RNA Seq	GSM1629099		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629099	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S1797.R1.fq zf3_dis_S1797.R2.fq	fastq fastq	15301850.0	306037.0	GSM1629099 r1	0:25 1:25	A:3854776;C:3654525;G:3713786;T:4018813;N:59950	25	25	3854776	3654525	3713786	4018813	59950	SRX914866	SRS869436	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.854	0.76358	0.30542	0.28721	0.87448	0.89392	0.62852	0.70188	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38485	38485	SRR1872973	SRX914865	SRS869437	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S1796	GSM1629098	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S1796	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629098	GSM1629098: zf3 dis S1796; Danio rerio; RNA Seq	GSM1629098		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629098	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S1796.R2.fq zf3_dis_S1796.R1.fq	fastq fastq	8775100.0	175502.0	GSM1629098 r1	0:25 1:25	A:2202984;C:2111992;G:2131355;T:2296008;N:32761	25	25	2202984	2111992	2131355	2296008	32761	SRX914865	SRS869437	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.73886	0.82202	0.24645	0.25948	0.89069	0.87282	0.71893	0.66123	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38486	38486	SRR1872972	SRX914864	SRS869438	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S1795	GSM1629097	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S1795	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629097	GSM1629097: zf3 dis S1795; Danio rerio; RNA Seq	GSM1629097		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629097	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S1795.R1.fq zf3_dis_S1795.R2.fq	fastq fastq	3867650.0	77353.0	GSM1629097 r1	0:25 1:25	A:945418;C:947115;G:963293;T:998039;N:13785	25	25	945418	947115	963293	998039	13785	SRX914864	SRS869438	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.84833	0.75285	0.27938	0.26663	0.91543	0.93026	0.63501	0.69265	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38487	38487	SRR1872971	SRX914863	SRS869406	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S1794	GSM1629096	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S1794	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629096	GSM1629096: zf3 dis S1794; Danio rerio; RNA Seq	GSM1629096		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629096	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S1794.R1.fq zf3_dis_S1794.R2.fq	fastq fastq	8452150.0	169043.0	GSM1629096 r1	0:25 1:25	A:2071197;C:2071760;G:2110479;T:2169192;N:29522	25	25	2071197	2071760	2110479	2169192	29522	SRX914863	SRS869406	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.73958	0.83971	0.24128	0.26353	0.91887	0.90023	0.72096	0.64372	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38488	38488	SRR1872970	SRX914862	SRS869439	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S1793	GSM1629095	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S1793	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629095	GSM1629095: zf3 dis S1793; Danio rerio; RNA Seq	GSM1629095		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629095	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S1793.R1.fq zf3_dis_S1793.R2.fq	fastq fastq	8832450.0	176649.0	GSM1629095 r1	0:25 1:25	A:2188095;C:2142691;G:2187563;T:2283127;N:30974	25	25	2188095	2142691	2187563	2283127	30974	SRX914862	SRS869439	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.82613	0.72679	0.22978	0.21996	0.89323	0.91263	0.60951	0.69076	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38489	38489	SRR1872969	SRX914861	SRS869405	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S1792	GSM1629094	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S1792	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629094	GSM1629094: zf3 dis S1792; Danio rerio; RNA Seq	GSM1629094		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629094	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S1792.R1.fq zf3_dis_S1792.R2.fq	fastq fastq	5363400.0	107268.0	GSM1629094 r1	0:25 1:25	A:1333155;C:1295485;G:1314864;T:1400923;N:18973	25	25	1333155	1295485	1314864	1400923	18973	SRX914861	SRS869405	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.79882	0.71843	0.19651	0.18206	0.91967	0.93679	0.67635	0.75194	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38490	38490	SRR1872968	SRX914860	SRS869440	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S1791	GSM1629093	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S1791	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629093	GSM1629093: zf3 dis S1791; Danio rerio; RNA Seq	GSM1629093		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629093	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S1791.R1.fq zf3_dis_S1791.R2.fq	fastq fastq	6287000.0	125740.0	GSM1629093 r1	0:25 1:25	A:1573148;C:1505013;G:1531762;T:1654056;N:23021	25	25	1573148	1505013	1531762	1654056	23021	SRX914860	SRS869440	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.7364	0.83315	0.21321	0.23072	0.92579	0.90887	0.74029	0.65374	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38491	38491	SRR1872967	SRX914859	SRS869441	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S1790	GSM1629092	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S1790	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629092	GSM1629092: zf3 dis S1790; Danio rerio; RNA Seq	GSM1629092		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629092	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S1790.R1.fq zf3_dis_S1790.R2.fq	fastq fastq	10112150.0	202243.0	GSM1629092 r1	0:25 1:25	A:2508260;C:2430588;G:2481266;T:2652826;N:39210	25	25	2508260	2430588	2481266	2652826	39210	SRX914859	SRS869441	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.76794	0.86354	0.24185	0.25375	0.90084	0.87858	0.7102	0.64235	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38492	38492	SRR1872966	SRX914858	SRS869402	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S1789	GSM1629091	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S1789	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629091	GSM1629091: zf3 dis S1789; Danio rerio; RNA Seq	GSM1629091		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629091	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S1789.R1.fq zf3_dis_S1789.R2.fq	fastq fastq	11231050.0	224621.0	GSM1629091 r1	0:25 1:25	A:2804694;C:2700654;G:2738351;T:2946403;N:40948	25	25	2804694	2700654	2738351	2946403	40948	SRX914858	SRS869402	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.77491	0.865	0.23972	0.24944	0.90173	0.88138	0.73586	0.39217	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38493	38493	SRR1872965	SRX914857	SRS869401	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S1788	GSM1629090	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S1788	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629090	GSM1629090: zf3 dis S1788; Danio rerio; RNA Seq	GSM1629090		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629090	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S1788.R1.fq zf3_dis_S1788.R2.fq	fastq fastq	10314050.0	206281.0	GSM1629090 r1	0:25 1:25	A:2552958;C:2507831;G:2553958;T:2660287;N:39016	25	25	2552958	2507831	2553958	2660287	39016	SRX914857	SRS869401	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.75288	0.84683	0.24666	0.25857	0.9178	0.89785	0.70009	0.64569	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38494	38494	SRR1872964	SRX914856	SRS869404	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S1787	GSM1629089	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S1787	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629089	GSM1629089: zf3 dis S1787; Danio rerio; RNA Seq	GSM1629089		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629089	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S1787.R1.fq zf3_dis_S1787.R2.fq	fastq fastq	10036550.0	200731.0	GSM1629089 r1	0:25 1:25	A:2476382;C:2435406;G:2478877;T:2611435;N:34450	25	25	2476382	2435406	2478877	2611435	34450	SRX914856	SRS869404	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.76416	0.85721	0.24235	0.25133	0.91019	0.89085	0.72942	0.65077	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38495	38495	SRR1872963	SRX914855	SRS869403	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S1786	GSM1629088	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S1786	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629088	GSM1629088: zf3 dis S1786; Danio rerio; RNA Seq	GSM1629088		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629088	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S1786.R1.fq zf3_dis_S1786.R2.fq	fastq fastq	10707900.0	214158.0	GSM1629088 r1	0:25 1:25	A:2626362;C:2606657;G:2669488;T:2766478;N:38915	25	25	2626362	2606657	2669488	2766478	38915	SRX914855	SRS869403	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.76097	0.85966	0.23352	0.2479	0.91352	0.89244	0.70793	0.63433	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38496	38496	SRR1872962	SRX914854	SRS869400	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S1785	GSM1629087	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S1785	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629087	GSM1629087: zf3 dis S1785; Danio rerio; RNA Seq	GSM1629087		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629087	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S1785.R1.fq zf3_dis_S1785.R2.fq	fastq fastq	13552600.0	271052.0	GSM1629087 r1	0:25 1:25	A:3331733;C:3298975;G:3371183;T:3504256;N:46453	25	25	3331733	3298975	3371183	3504256	46453	SRX914854	SRS869400	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.8397	0.74277	0.22428	0.21707	0.88653	0.90569	0.6195	0.69908	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38497	38497	SRR1872961	SRX914853	SRS869443	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S1784	GSM1629086	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S1784	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629086	GSM1629086: zf3 dis S1784; Danio rerio; RNA Seq	GSM1629086		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629086	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S1784.R1.fq zf3_dis_S1784.R2.fq	fastq fastq	4404200.0	88084.0	GSM1629086 r1	0:25 1:25	A:1084888;C:1067528;G:1088523;T:1145020;N:18241	25	25	1084888	1067528	1088523	1145020	18241	SRX914853	SRS869443	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.8337	0.73219	0.2522	0.22849	0.92415	0.93973	0.64174	0.72344	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38498	38498	SRR1872960	SRX914852	SRS869446	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S1783	GSM1629085	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S1783	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629085	GSM1629085: zf3 dis S1783; Danio rerio; RNA Seq	GSM1629085		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629085	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S1783.R1.fq zf3_dis_S1783.R2.fq	fastq fastq	8078100.0	161562.0	GSM1629085 r1	0:25 1:25	A:2018080;C:1941208;G:1967979;T:2119661;N:31172	25	25	2018080	1941208	1967979	2119661	31172	SRX914852	SRS869446	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.75378	0.84405	0.23097	0.24639	0.9249	0.90733	0.72402	0.6512	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38499	38499	SRR1872959	SRX914851	SRS869447	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S1782	GSM1629084	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S1782	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629084	GSM1629084: zf3 dis S1782; Danio rerio; RNA Seq	GSM1629084		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629084	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S1782.R1.fq zf3_dis_S1782.R2.fq	fastq fastq	11804850.0	236097.0	GSM1629084 r1	0:25 1:25	A:2917509;C:2862283;G:2913328;T:3066429;N:45301	25	25	2917509	2862283	2913328	3066429	45301	SRX914851	SRS869447	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.7534	0.85203	0.23766	0.25262	0.90946	0.88763	0.696	0.62863	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38500	38500	SRR1872958	SRX914850	SRS869444	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S1781	GSM1629083	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S1781	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629083	GSM1629083: zf3 dis S1781; Danio rerio; RNA Seq	GSM1629083		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629083	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S1781.R1.fq zf3_dis_S1781.R2.fq	fastq fastq	19207550.0	384151.0	GSM1629083 r1	0:25 1:25	A:4696459;C:4696876;G:4796915;T:4950373;N:66927	25	25	4696459	4696876	4796915	4950373	66927	SRX914850	SRS869444	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.8568	0.76431	0.29961	0.28449	0.88538	0.90412	0.65058	0.72551	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38501	38501	SRR1872957	SRX914849	SRS869442	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S1780	GSM1629082	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S1780	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629082	GSM1629082: zf3 dis S1780; Danio rerio; RNA Seq	GSM1629082		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629082	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S1780.R1.fq zf3_dis_S1780.R2.fq	fastq fastq	11145800.0	222916.0	GSM1629082 r1	0:25 1:25	A:2786859;C:2681494;G:2712839;T:2922888;N:41720	25	25	2786859	2681494	2712839	2922888	41720	SRX914849	SRS869442	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.83046	0.74858	0.26498	0.24912	0.8899	0.90974	0.64495	0.72691	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38502	38502	SRR1872956	SRX914848	SRS869445	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S1779	GSM1629081	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S1779	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629081	GSM1629081: zf3 dis S1779; Danio rerio; RNA Seq	GSM1629081		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629081	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S1779.R1.fq zf3_dis_S1779.R2.fq	fastq fastq	9910000.0	198200.0	GSM1629081 r1	0:25 1:25	A:2571262;C:2341911;G:2362251;T:2600524;N:34052	25	25	2571262	2341911	2362251	2600524	34052	SRX914848	SRS869445	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.73701	0.81634	0.22371	0.2248	0.9237	0.90485	0.71858	0.63582	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38503	38503	SRR1872955	SRX914847	SRS869448	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S1778	GSM1629080	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S1778	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629080	GSM1629080: zf3 dis S1778; Danio rerio; RNA Seq	GSM1629080		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629080	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S1778.R1.fq zf3_dis_S1778.R2.fq	fastq fastq	7057350.0	141147.0	GSM1629080 r1	0:25 1:25	A:1744846;C:1725519;G:1755935;T:1806453;N:24597	25	25	1744846	1725519	1755935	1806453	24597	SRX914847	SRS869448	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.7354	0.82457	0.24353	0.25629	0.93074	0.91522	0.70619	0.61979	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38504	38504	SRR1872954	SRX914846	SRS869450	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S1777	GSM1629079	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S1777	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629079	GSM1629079: zf3 dis S1777; Danio rerio; RNA Seq	GSM1629079		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629079	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S1777.R2.fq zf3_dis_S1777.R1.fq	fastq fastq	5846650.0	116933.0	GSM1629079 r1	0:25 1:25	A:1452630;C:1418080;G:1437872;T:1517380;N:20688	25	25	1452630	1418080	1437872	1517380	20688	SRX914846	SRS869450	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.80248	0.70977	0.23508	0.22075	0.92857	0.94235	0.62387	0.69869	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38505	38505	SRR1872953	SRX914845	SRS869449	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S1776	GSM1629078	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S1776	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629078	GSM1629078: zf3 dis S1776; Danio rerio; RNA Seq	GSM1629078		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629078	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S1776.R1.fq zf3_dis_S1776.R2.fq	fastq fastq	31020450.0	620409.0	GSM1629078 r1	0:25 1:25	A:7743857;C:7444773;G:7480098;T:8231346;N:120376	25	25	7743857	7444773	7480098	8231346	120376	SRX914845	SRS869449	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.86304	0.7617	0.25255	0.23195	0.87507	0.89629	0.6564	0.73758	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38506	38506	SRR1872952	SRX914844	SRS869453	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S1775	GSM1629077	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S1775	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629077	GSM1629077: zf3 dis S1775; Danio rerio; RNA Seq	GSM1629077		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629077	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S1775.R1.fq zf3_dis_S1775.R2.fq	fastq fastq	26769150.0	535383.0	GSM1629077 r1	0:25 1:25	A:6747957;C:6390810;G:6390320;T:7139562;N:100501	25	25	6747957	6390810	6390320	7139562	100501	SRX914844	SRS869453	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.85207	0.7656	0.23893	0.22295	0.89148	0.91185	0.67617	0.75817	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38507	38507	SRR1872951	SRX914843	SRS869399	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S1774	GSM1629076	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S1774	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629076	GSM1629076: zf3 dis S1774; Danio rerio; RNA Seq	GSM1629076		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629076	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S1774.R1.fq zf3_dis_S1774.R2.fq	fastq fastq	26813750.0	536275.0	GSM1629076 r1	0:25 1:25	A:6794513;C:6357376;G:6372394;T:7190805;N:98662	25	25	6794513	6357376	6372394	7190805	98662	SRX914843	SRS869399	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.78235	0.86464	0.22966	0.24793	0.91569	0.89721	0.77946	0.708	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38508	38508	SRR1872950	SRX914842	SRS869451	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S1773	GSM1629075	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S1773	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629075	GSM1629075: zf3 dis S1773; Danio rerio; RNA Seq	GSM1629075		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629075	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S1773.R1.fq zf3_dis_S1773.R2.fq	fastq fastq	25089650.0	501793.0	GSM1629075 r1	0:25 1:25	A:6132687;C:6129998;G:6227748;T:6528593;N:70624	25	25	6132687	6129998	6227748	6528593	70624	SRX914842	SRS869451	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.76552	0.86499	0.24948	0.26604	0.85141	0.83155	0.72704	0.64381	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38509	38509	SRR1872949	SRX914841	SRS869452	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S1772	GSM1629074	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S1772	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629074	GSM1629074: zf3 dis S1772; Danio rerio; RNA Seq	GSM1629074		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629074	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S1772.R1.fq zf3_dis_S1772.R2.fq	fastq fastq	14452950.0	289059.0	GSM1629074 r1	0:25 1:25	A:3398203;C:3659341;G:3644357;T:3705062;N:45987	25	25	3398203	3659341	3644357	3705062	45987	SRX914841	SRS869452	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.85023	0.79726	0.37094	0.34499	0.95256	0.96122	0.7469	0.8274	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38510	38510	SRR1872948	SRX914840	SRS869454	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S1771	GSM1629073	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S1771	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629073	GSM1629073: zf3 dis S1771; Danio rerio; RNA Seq	GSM1629073		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629073	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S1771.R1.fq zf3_dis_S1771.R2.fq	fastq fastq	28010450.0	560209.0	GSM1629073 r1	0:25 1:25	A:7013721;C:6726949;G:6738187;T:7426108;N:105485	25	25	7013721	6726949	6738187	7426108	105485	SRX914840	SRS869454	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.8442	0.76206	0.25757	0.24708	0.82495	0.84382	0.66364	0.74618	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38511	38511	SRR1872947	SRX914839	SRS869455	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S1770	GSM1629072	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S1770	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629072	GSM1629072: zf3 dis S1770; Danio rerio; RNA Seq	GSM1629072		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629072	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S1770.R1.fq zf3_dis_S1770.R2.fq	fastq fastq	26047350.0	520947.0	GSM1629072 r1	0:25 1:25	A:6460451;C:6292275;G:6355799;T:6847889;N:90936	25	25	6460451	6292275	6355799	6847889	90936	SRX914839	SRS869455	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.86221	0.76674	0.24438	0.23105	0.89404	0.91392	0.65574	0.72729	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38512	38512	SRR1872946	SRX914838	SRS869456	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S1769	GSM1629071	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S1769	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629071	GSM1629071: zf3 dis S1769; Danio rerio; RNA Seq	GSM1629071		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629071	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S1769.R1.fq zf3_dis_S1769.R2.fq	fastq fastq	23466650.0	469333.0	GSM1629071 r1	0:25 1:25	A:5650465;C:5821792;G:5968440;T:5964599;N:61354	25	25	5650465	5821792	5968440	5964599	61354	SRX914838	SRS869456	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.85558	0.75262	0.26105	0.24853	0.81899	0.83767	0.6086	0.68957	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38513	38513	SRR1872945	SRX914837	SRS869457	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S1768	GSM1629070	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S1768	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629070	GSM1629070: zf3 dis S1768; Danio rerio; RNA Seq	GSM1629070		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629070	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S1768.R1.fq zf3_dis_S1768.R2.fq	fastq fastq	34233750.0	684675.0	GSM1629070 r1	0:25 1:25	A:8552131;C:8192670;G:8299965;T:9064745;N:124239	25	25	8552131	8192670	8299965	9064745	124239	SRX914837	SRS869457	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.77382	0.86348	0.23973	0.25388	0.89714	0.87663	0.74271	0.66277	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38514	38514	SRR1872944	SRX914836	SRS869459	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S1767	GSM1629069	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S1767	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629069	GSM1629069: zf3 dis S1767; Danio rerio; RNA Seq	GSM1629069		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629069	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S1767.R1.fq zf3_dis_S1767.R2.fq	fastq fastq	41012700.0	820254.0	GSM1629069 r1	0:25 1:25	A:10279155;C:9845951;G:9891678;T:10856002;N:139914	25	25	10279155	9845951	9891678	10856002	139914	SRX914836	SRS869459	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.73067	0.81337	0.2169	0.22525	0.88323	0.85934	0.73824	0.39021	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38515	38515	SRR1872943	SRX914835	SRS869458	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S1766	GSM1629068	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S1766	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629068	GSM1629068: zf3 dis S1766; Danio rerio; RNA Seq	GSM1629068		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629068	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S1766.R1.fq zf3_dis_S1766.R2.fq	fastq fastq	29188250.0	583765.0	GSM1629068 r1	0:25 1:25	A:7242191;C:7031890;G:7076968;T:7731974;N:105227	25	25	7242191	7031890	7076968	7731974	105227	SRX914835	SRS869458	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.78501	0.86164	0.25669	0.26945	0.9065	0.88631	0.76335	0.68598	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38516	38516	SRR1872942	SRX914834	SRS869460	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S1765	GSM1629067	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S1765	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629067	GSM1629067: zf3 dis S1765; Danio rerio; RNA Seq	GSM1629067		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629067	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S1765.R1.fq zf3_dis_S1765.R2.fq	fastq fastq	20964300.0	419286.0	GSM1629067 r1	0:25 1:25	A:5197796;C:5061573;G:5144805;T:5487948;N:72178	25	25	5197796	5061573	5144805	5487948	72178	SRX914834	SRS869460	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.75339	0.84553	0.23903	0.25244	0.90889	0.88943	0.72905	0.65459	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38517	38517	SRR1872941	SRX914833	SRS869461	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S1764	GSM1629066	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S1764	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629066	GSM1629066: zf3 dis S1764; Danio rerio; RNA Seq	GSM1629066		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629066	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S1764.R1.fq zf3_dis_S1764.R2.fq	fastq fastq	19622100.0	392442.0	GSM1629066 r1	0:25 1:25	A:4858560;C:4737576;G:4827858;T:5126022;N:72084	25	25	4858560	4737576	4827858	5126022	72084	SRX914833	SRS869461	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.85664	0.76261	0.26266	0.24985	0.89215	0.9121	0.64591	0.72137	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38518	38518	SRR1872940	SRX914832	SRS869462	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S1763	GSM1629065	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S1763	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629065	GSM1629065: zf3 dis S1763; Danio rerio; RNA Seq	GSM1629065		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629065	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S1763.R1.fq zf3_dis_S1763.R2.fq	fastq fastq	25578050.0	511561.0	GSM1629065 r1	0:25 1:25	A:6428474;C:6113953;G:6157283;T:6780850;N:97490	25	25	6428474	6113953	6157283	6780850	97490	SRX914832	SRS869462	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.76356	0.85195	0.24512	0.25395	0.89613	0.8759	0.72474	0.65036	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38519	38519	SRR1872939	SRX914831	SRS869463	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S1762	GSM1629064	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S1762	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629064	GSM1629064: zf3 dis S1762; Danio rerio; RNA Seq	GSM1629064		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629064	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S1762.R1.fq zf3_dis_S1762.R2.fq	fastq fastq	1653650.0	33073.0	GSM1629064 r1	0:25 1:25	A:583314;C:332778;G:308607;T:423239;N:5712	25	25	583314	332778	308607	423239	5712	SRX914831	SRS869463	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.49477	0.45686	0.22692	0.19579	0.99007	0.98845	0.78316	0.72207	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38520	38520	SRR1872938	SRX914830	SRS869464	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S1761	GSM1629063	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S1761	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629063	GSM1629063: zf3 dis S1761; Danio rerio; RNA Seq	GSM1629063		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629063	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S1761.R2.fq zf3_dis_S1761.R1.fq	fastq fastq	19984900.0	399698.0	GSM1629063 r1	0:25 1:25	A:4961479;C:4806885;G:4904814;T:5240574;N:71148	25	25	4961479	4806885	4904814	5240574	71148	SRX914830	SRS869464	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.75392	0.84898	0.22315	0.24029	0.91431	0.89445	0.7209	0.64061	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38521	38521	SRR1872937	SRX914829	SRS869468	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S1760	GSM1629062	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S1760	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629062	GSM1629062: zf3 dis S1760; Danio rerio; RNA Seq	GSM1629062		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629062	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S1760.R1.fq zf3_dis_S1760.R2.fq	fastq fastq	22454850.0	449097.0	GSM1629062 r1	0:25 1:25	A:5586507;C:5388486;G:5444119;T:5950465;N:85273	25	25	5586507	5388486	5444119	5950465	85273	SRX914829	SRS869468	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.7333	0.82243	0.23349	0.24842	0.92269	0.90167	0.74655	0.66989	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38522	38522	SRR1872936	SRX914828	SRS869465	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S1759	GSM1629061	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S1759	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629061	GSM1629061: zf3 dis S1759; Danio rerio; RNA Seq	GSM1629061		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629061	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S1759.R1.fq zf3_dis_S1759.R2.fq	fastq fastq	27417700.0	548354.0	GSM1629061 r1	0:25 1:25	A:6897583;C:6528103;G:6564835;T:7326040;N:101139	25	25	6897583	6528103	6564835	7326040	101139	SRX914828	SRS869465	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.86104	0.77337	0.23833	0.2267	0.88503	0.90682	0.67192	0.75232	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38523	38523	SRR1872935	SRX914827	SRS869467	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S1758	GSM1629060	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S1758	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629060	GSM1629060: zf3 dis S1758; Danio rerio; RNA Seq	GSM1629060		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629060	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S1758.R1.fq zf3_dis_S1758.R2.fq	fastq fastq	30382900.0	607658.0	GSM1629060 r1	0:25 1:25	A:7574204;C:7302169;G:7332368;T:8062062;N:112097	25	25	7574204	7302169	7332368	8062062	112097	SRX914827	SRS869467	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.77207	0.8647	0.23912	0.25034	0.90666	0.88586	0.73416	0.65776	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38524	38524	SRR1872934	SRX914826	SRS869469	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S1757	GSM1629059	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S1757	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629059	GSM1629059: zf3 dis S1757; Danio rerio; RNA Seq	GSM1629059		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629059	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S1757.R2.fq zf3_dis_S1757.R1.fq	fastq fastq	29399850.0	587997.0	GSM1629059 r1	0:25 1:25	A:7302988;C:7069568;G:7111055;T:7812507;N:103732	25	25	7302988	7069568	7111055	7812507	103732	SRX914826	SRS869469	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.76741	0.86323	0.22734	0.23778	0.89623	0.87428	0.744	0.65096	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38525	38525	SRR1872933	SRX914825	SRS869466	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S1756	GSM1629058	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S1756	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629058	GSM1629058: zf3 dis S1756; Danio rerio; RNA Seq	GSM1629058		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629058	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S1756.R2.fq zf3_dis_S1756.R1.fq	fastq fastq	28316550.0	566331.0	GSM1629058 r1	0:25 1:25	A:6939851;C:6914142;G:6969668;T:7393905;N:98984	25	25	6939851	6914142	6969668	7393905	98984	SRX914825	SRS869466	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.75503	0.84546	0.22273	0.22989	0.91208	0.89215	0.74047	0.6539	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38526	38526	SRR1872932	SRX914824	SRS869470	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S1755	GSM1629057	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S1755	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629057	GSM1629057: zf3 dis S1755; Danio rerio; RNA Seq	GSM1629057		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629057	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S1755.R1.fq zf3_dis_S1755.R2.fq	fastq fastq	34522150.0	690443.0	GSM1629057 r1	0:25 1:25	A:8664696;C:8269609;G:8245763;T:9216755;N:125327	25	25	8664696	8269609	8245763	9216755	125327	SRX914824	SRS869470	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.70965	0.79198	0.20306	0.20735	0.8997	0.87742	0.73652	0.65101	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38527	38527	SRR1872931	SRX914823	SRS869471	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S1754	GSM1629056	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S1754	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629056	GSM1629056: zf3 dis S1754; Danio rerio; RNA Seq	GSM1629056		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629056	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S1754.R1.fq zf3_dis_S1754.R2.fq	fastq fastq	2516450.0	50329.0	GSM1629056 r1	0:25 1:25	A:733640;C:576645;G:555236;T:641989;N:8940	25	25	733640	576645	555236	641989	8940	SRX914823	SRS869471	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.52845	0.57434	0.22207	0.24897	0.98474	0.98159	0.75341	0.69085	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38528	38528	SRR1872930	SRX914822	SRS869397	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S1753	GSM1629055	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S1753	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629055	GSM1629055: zf3 dis S1753; Danio rerio; RNA Seq	GSM1629055		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629055	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S1753.R2.fq zf3_dis_S1753.R1.fq	fastq fastq	28269750.0	565395.0	GSM1629055 r1	0:25 1:25	A:7018818;C:6799965;G:6877890;T:7473678;N:99399	25	25	7018818	6799965	6877890	7473678	99399	SRX914822	SRS869397	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.76795	0.8582	0.22532	0.23281	0.88323	0.85999	0.72315	0.63966	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures
38529	38529	SRR1872929	SRX914821	SRS869398	SRP055996	PRJNA277737	Spatial reconstruction of single cell gene expression	GSE66688	Transcriptome Analysis	Spatial localization is a key determinant of cellular fate and behavior but spatial RNA assays traditionally rely on staining for a limited number of RNA species. In contrast single cell RNA seq allows for deep profiling of cellular gene expression but established methods separate cells from their native spatial context. Here we present Seurat a computational strategy to infer cellular localization by integrating single cell RNA seq data with in situ RNA patterns. We applied Seurat to spatially map 851 single cells from dissociated zebrafish Danio rerio embryos inferring a transcriptome wide map of spatial patterning. We confirmed Seurat’s accuracy using several experimental approaches and used it to identify a set of archetypal expression patterns and spatial markers. Additionally Seurat correctly localizes rare subpopulations accurately mapping both spatially restricted and scattered groups. Seurat will be applicable to mapping cellular localization within complex patterned tissues in diverse systems. Overall design: We generated single cell RNA seq profiles from dissociated cells from developing zebrafish embryos late blastula stage 50% epiboly	pubmed:25867923	zf3 dis S1752	GSM1629054	tissue:50% Epiboly stage\|group:experimental batch 3	zf3 dis S1752	To map reads we slightly modified the Zv9 reference transcriptome since our reads were expected to originate from the 5’ end of each mRNA molecule which could cause problems if the reference gene models had an incorrect annotation for the transcription start site TSS. To address this we extended the TSS for all Zv9 genes by 100 bases upstream to allow for minor fluctuations. We then aligned read pairs directly to this modified reference transcriptome using Bowtie with the following parameters: q phred33 quals n 2 l 25 I 1 X 2000 a m 200. As expected following mapping our reads overwhelmingly originated from near the TSS. To quantify gene expression we leveraged the 5 bp Random Molecular Tags RMTs that were associated with each sequencing read pair. For each annotated gene in Zv9 we identified all read pairs that mapped to the correct strand and collected each of the 5 bp RMTs that were associated with these reads. We collapsed duplicate RMT sequences together in order to calculate the number of distinct RMTs associated with each gene. We quantified gene expression for 1 152 single cell libraries and identified a subset of 207 failed/low quality libraries with poor transcriptome complexity < 2 000 genes detected per cell. post excluding these remaining dataset of 945 cells on average we identified 47 000 unique molecules per sequencing library corresponding to the detection of 3 400 genes per single cell. Supplementary files format and content: zdata.expMatrix.txt Expression matrix for all samples	50% Epiboly stage		Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	Fertilized eggs from TL/AB in crosses were incubated in 1 mg/ml pronase Protease from Streptomyces griseus Sigma Aldrich in a glass dish for 4–5 minutes until the chorion began to blister. The embryos were submerged in 200 ml of embryo medium in a glass beaker without xxx them to contact air or plastic both of which will cause the embryos to burst. The medium was poured off and new medium was vigorously added again without xxx embryos to contact air twice in order to mechanically remove the chorions. The embryos were cultured in petri dishes coated with 2% agarose to avoid contact with plastic either at 23°C or 28°C until they reached 50% epiboly about 6 hpf at 28°C. At 50% epiboly single embryos were visually confirmed to be at the correct stage	group:experimental batch 3	GSM1629054	GSM1629054: zf3 dis S1752; Danio rerio; RNA Seq	GSM1629054		1	Two pairs of watchmaker forceps were used to dissect the blastula cap of the embryo away from the yolk. First one pair of forceps was used to hold and rotate the cap while the other was used to cut and pinch away the yolk that extended below the blastula cap. Then the blastula cap was cut slightly up the side which exposed the yolk that was inside of the blastula cap which could then be gently peeled away. The blastula cap was transferred by pipette into a microfuge tube that contained 60 μl of DMEM/F12 media. The cells were dissociated by vigorously flicking the tube 10 times and then pipetting the entire volume twice while visually confirming that dissociation had occurred. A timer was started at this time to track the amount of time the embryo had been dissociated. If cell clumps were still visible the tube was flicked again. 180 μl of DMEM/F12 was added to dilute the cell mixture then 120 μl of the diluted cell mixture was pipetted across the surface of a new agarose coated dish filled with DMEM/F12 media. To collect cells a P2 pipettor was used while observing the cells under the dissecting scope to collect 0.5 μl of media that contained a single cell. This was pipetted into 3 μl of TCL lysis buffer Qiagen in the lid of a PCR strip and mixed 3 times to ensure that lysis occurred. post collection of 8 cells the entire strip of lids was snapped into a 96 well plate and kept on dry ice. We used a modified strand specific single cell RNA seq protocol based on the SMART template switching method where the template switch oligonucleotide included a stretch of 5 randomized nucleotides thereby tagging each mRNA molecule with a random molecular tag RMT prior to PCR to mitigate amplification bias. Furthermore we used a modified library preparation protocol that shares similarities with Soumillon et al. and is based on the Nextera Sample Preparation Kit. It selectively amplifies the 5’ transcript end retains strand information and is compatible with standard Illumina sequencing primers.	GEO Accession:GSM1629054	RNA-Seq	TRANSCRIPTOMIC	cDNA	PAIRED	ILLUMINA	Illumina HiSeq 2500	SRP055996	loader:fastq load.py	zf3_dis_S1752.R2.fq zf3_dis_S1752.R1.fq	fastq fastq	28368800.0	567376.0	GSM1629054 r1	0:25 1:25	A:7064616;C:6791290;G:6863490;T:7546661;N:102743	25	25	7064616	6791290	6863490	7546661	102743	SRX914821	SRS869398	SRA246089	GEO	Satija Lab, New York Genome Center	2	0.86602	0.76935	0.24904	0.23406	0.88538	0.9067	0.65965	0.73571	25	25	B	B	biological fallback assumption	illumina	hiseq_era	unknown	poly_a	nextera	sc	single_cell_plate	smartseq	United States	2015-03-09	Multi-stage	Embryo	Embryo Imprecise	All anatomical structures

Advanced export

JSON shape: default, array, newline-delimited

CREATE TABLE run_metadata("run.accession" VARCHAR, "experiment.accession" VARCHAR, "sample.accession" VARCHAR, "study.accession" VARCHAR, bioproject VARCHAR, "study.title" VARCHAR, "study.alias" VARCHAR, "study.type" VARCHAR, "study.abstract" VARCHAR, "study.attributes" VARCHAR, "study.PMIDs" VARCHAR, "sample.description" VARCHAR, "sample.title" VARCHAR, "sample.alias" VARCHAR, "sample.centername" VARCHAR, "sample.attributes" VARCHAR, "GEOsample.title" VARCHAR, "GEOsample.dataprocessing" VARCHAR, "GEOsample.source" VARCHAR, "GEOsample.treatmentprotocol" VARCHAR, "GEOsample.extractprotocol" VARCHAR, "GEOsample.growthprotocol" VARCHAR, "GEOsample.characteristics" VARCHAR, "GEOsample.accession" VARCHAR, "experiment.title" VARCHAR, "experiment.alias" VARCHAR, "experiment.library_name" VARCHAR, "experiment.design_description" VARCHAR, "experiment.library_construction_protocol" VARCHAR, "experiment.attributes" VARCHAR, "experiment.library_strategy" VARCHAR, "experiment.library_source" VARCHAR, "experiment.library_selection" VARCHAR, "experiment.library_layout" VARCHAR, "experiment.platform" VARCHAR, "experiment.instrument_model" VARCHAR, "experiment.spot_descriptor" VARCHAR, "experiment.study_ref" VARCHAR, "run.title" VARCHAR, "run.attributes" VARCHAR, "run.filename" VARCHAR, "run.semantic_name" VARCHAR, "run.total_bases" DOUBLE, "run.total_spots" DOUBLE, "run.alias" VARCHAR, "run.read_lengths" VARCHAR, "run.base_counts" VARCHAR, "run.r1_length" BIGINT, "run.r2_length" BIGINT, "run.r3_length" BIGINT, "run.r4_length" BIGINT, "run.Acount" BIGINT, "run.Ccount" BIGINT, "run.Gcount" BIGINT, "run.Tcount" BIGINT, "run.Ncount" BIGINT, "run.experiment" VARCHAR, "run.pool_member" VARCHAR, "submission.accession" VARCHAR, "submission.srasource" VARCHAR, "submission.bioprojectsource" VARCHAR, "seqdetective.n_mates" BIGINT, "seqdetective.mapping_rate.mate1" DOUBLE, "seqdetective.mapping_rate.mate2" DOUBLE, "seqdetective.nofeature_rate.mate1" DOUBLE, "seqdetective.nofeature_rate.mate2" DOUBLE, "seqdetective.sparsity.mate1" DOUBLE, "seqdetective.sparsity.mate2" DOUBLE, "seqdetective.pos_strand_rate.mate1" DOUBLE, "seqdetective.pos_strand_rate.mate2" DOUBLE, "seqdetective.readlen.mate1" BIGINT, "seqdetective.readlen.mate2" BIGINT, "seqdetective.judgement.mate1" VARCHAR, "seqdetective.judgement.mate2" VARCHAR, "seqdetective.judgement.reason" VARCHAR, platform_family VARCHAR, instrument_generation VARCHAR, read_bias VARCHAR, selection_class VARCHAR, prep_kit VARCHAR, sc_or_bulk VARCHAR, tech_class VARCHAR, technology VARCHAR, tech_variant VARCHAR, "submission.bioprojectsource.country" VARCHAR, earliest_date DATE, devstage_curation VARCHAR, devstage_curation_coarse VARCHAR, tissue_curation VARCHAR, tissue_curation_coarse VARCHAR);;
CREATE INDEX idx_run_bioproject ON run_metadata(bioproject);;
CREATE INDEX idx_run_run_accession ON run_metadata("run.accession");;