Integrated analysis of single-cell embryo data yields a unified transcriptome signature for the human preimplantation epiblast
Giuliano Giuseppe StirparoiD, Thorsten BoroviakiD, Ge GuoiD, Jennifer NicholsiD, Austin G SmithiD, Paul BertoneiD
Abstract
Single-cell profiling techniques create opportunities to delineate cell fate progression in mammalian development. Recent studies provide transcriptome data from human preimplantation embryos, in total comprising nearly 2000 individual cells. Interpretation of these data is confounded by biological factors such as variable embryo staging and cell-type ambiguity, as well as technical challenges in the collective analysis of datasets produced with different sample preparation and sequencing protocols. Here we address these issues to assemble a complete gene expression time course spanning human preimplantation embryogenesis. We identify key transcriptional features over developmental time and elucidate lineage-specific regulatory networks. We resolve post hoc cell-type assignment in the blastocyst, and define robust transcriptional prototypes that capture epiblast and primitive endoderm lineages. Examination of human pluripotent stem cell transcriptomes in this framework identifies culture conditions that sustain a naive state pertaining to the inner cell mass. Our approach thus clarifies understanding both of lineage segregation in the early human embryo and of in vitro stem cell identity, and provides an analytical resource for comparative molecular embryology.
02 /Specification & Classification
Click any tag to browse matching papers- Citations
- 205
- FWCI
- 9.78
- Field percentile
- top 1%
- Organism
- Assay
- Research Focus
- Approach
In-depth summary
Claude Opus 5.5 · xhigh effort · 8 Nov 2026 · from the full text of version 1 (21 Nov 2017) · CC BY 4.0
Reanalysis of three human embryo scRNA-seq datasets resolves cell-type ambiguities, defines epiblast and primitive endoderm prototypes, and identifies t2iLGö and 5i/L/A naïve stem cells as epiblast-like.
Three single-cell RNA-seq studies of human embryos (1683 transcriptomes) differ in staging, sample handling, sequencing protocol and cell labels, so epiblast (EPI) and primitive endoderm (PrE) cells were hard to tell apart consistently. This work merges the three datasets into one curated lineage reference and uses it to test which pluripotent stem cell (PSC) culture conditions best match the preimplantation epiblast.
Key results
ShownIn the largest dataset (Petropoulos et al.), cells labelled EPI and PrE overlap. Many cells collected by immunosurgery carry trophectoderm (TE) signatures, which points to incomplete TE removal and blurs the separation of lineages.
Immunosurgery cells came from four E6 and six E7 embryos, and over half were labelled TE in the original study. PCA on 1131 variable genes gave no distinct EPI/PrE groups; the genes driving separation were TE-associated (GATA2, GATA3, KRT8, DAB2, TEAD3). Labelled EPI and PrE also overlapped in t-SNE and diffusion maps. · TE overrepresentation in single-cell embryo data; Figure legends
ShownTwo steps let the authors pull late inner cell mass (ICM) cells out of the Petropoulos data: intersecting each study's variable genes, and selecting cells with high or medium POU5F1. These cells were merged with Yan and Blakeley cells into curated EPI, PrE and intermediate groups.
12 literature markers (four per lineage) separated Yan and Blakeley cells, matching the Blakeley labels. Intersecting per-study variable genes gave 188 shared genes and reduced study bias in PCA. Among 858 late blastocyst cells (1529 total), POU5F1-high/medium clusters grouped with Yan/Blakeley EPI and PrE, forming EPI, PrE and intermediate groups. · Lineage markers defining human EPI, PrE and TE; Representative transcriptomes of mature EPI and…; Figure legends
SuggestedAmong E5 cells, a separate early ICM state without EPI, PrE or TE signatures was found. It expresses ESRRB, a gene reported as early-ICM-specific in macaque and marmoset, and gene co-expression network analysis recovered it independently.
Pseudotime plus clustering isolated 43 of 336 E5 cells; ESRRB, ATG2A and MAGEA4 were high there and lower in EPI. WGCNA found 14 modules, merged into seven, whose eigengene clusters matched early ICM, EPI and PrE. Hub genes included MFN1, CYP26A1, NANOGNB (early ICM); PDGFRA, BMP6, RSPO3 (PrE); and GDF3, NANOG (EPI). · Pseudotime staging and non-human primate-derived…; Evaluation of refined embryonic cell populations; Figure legends
ShownA single expression map from zygote to late blastocyst yields stage-specific regulators and separate early and late marker sets for EPI and PrE. Agonists of FGF, Nodal and WNT signalling are restricted to EPI.
Self-organising maps (30×30 hexagonal) flagged LEUTX and KLF17 at the 8-cell stage and ZNF296 in compacted morula. ARGFX, PRDM14, SOX2, NANOG and DPPA2 start at genome activation and persist in EPI; late EPI adds FGF4, NODAL, LEFTY2, WNT3; early PrE shows GATA6, LAMA1, HNF4A; RSPO3, GATA4, SOX17, FOXA2 appear de novo in mature PrE. · A unified transcription map of human…; Identification of early and late preimplantation…
ShownNaïve PSC grown in t2iLGö (transgene-reset, chemically reset or embryo-derived) or in 5i/L/A most resemble preimplantation EPI. Cells grown in NHSM/4i, cells derived with HDAC inhibitors, and extended pluripotent stem cells group with conventional PSC.
By quadratic programming, the fraction of identity with EPI exceeded 0.75 for naïve cells cultured on laminin vs <0.54 for conventional and alternative PSC. On 2860 genes differentially expressed between naïve and conventional PSC, Pearson correlation with EPI was 0.66–0.8 (naïve) vs 0.19–0.44 (conventional); 5i/L/A reached 0.8 with both EPI and PrE. · Characteristics of human pluripotent stem cell…; Discussion; Figure legends
Methods, data and reproducibility
Public single-cell and bulk RNA-seq reads were realigned to GRCh38 with STAR and quantified the same way. Cells were reclassified using variable genes shared across studies, 12 lineage markers, POU5F1-based hierarchical clustering, pseudotime and WGCNA; PSC cultures were then compared with embryo stages by PCA, correlation and quadratic programming.
- Public single-cell RNA-seq of human preimplantation embryos from the Yan, Blakeley and Petropoulos studies, zygote to late blastocyst · human · n = 1683 transcriptomes (n=124; n=30; n=1529) · SRP011546; SRP055810; ERP012552
- Petropoulos et al. late blastocyst (E6–E7) cells, including an immunosurgery subset, used to isolate late ICM cells with high or medium POU5F1 · human · n = 858 late blastocyst cells of 1529
- Petropoulos et al. early blastocyst (E5) cells used to define an early ICM population · human · n = 336 E5 cells (43 early ICM)
- Conventional H1 and H9 PSC lines in standard culture · human · SRP014320; ERP007180
- Naïve PSC in t2iLGö: transgene-reset H9, chemically reset Shef6, embryo-derived HNES and HNES1 cultures, with conventional counterparts · human · ERP006823; ERP022538; ERP014247
- WIBR PSC lines reset to naïve pluripotency in 5i/L/A medium · human · SRP059227
- PSC in NHSM/4i (WIS, H1, Lis1) plus conventional H1 and Elf1 cultures · human · SRP045294; SRP045911
- Extended pluripotent stem (EPS) cells · human · SRP074076
- Methods & tools
- STAR 2.5.2b, htseq-count, DESeq2, FactoMineR, Rtsne, destiny (diffusion maps), scde (differential expression), DeconRNASeq (quadratic programming), WGCNA, kohonen (self-organising maps), GOstats; DAVID 6.8, Cytoscape with enrichment map plugin
- Data & accessions
- https://app.stemcells.cam.ac.uk/human-embryo
- Reagents & environment
- STAR 2.5.2b two-pass alignment to GRCh38/hg38; GENCODE v25 (Ensembl release 87) annotation; htseq-count; DAVID 6.8; AnimalTFDB 2.0; R/Bioconductor packages (DESeq2, scde, WGCNA, kohonen, DeconRNASeq), versions not stated.
- Reproducibility gaps
- No code repository is given, and the analyses relied partly on custom scripts. Processed expression data are provided only through Supplemental Tables and a web viewer for single genes. Most R package versions are not specified.
Limitations (6)
- Human IVF embryos cannot be staged precisely, so E5 cells showing lineage markers might come from more advanced embryos. The authors also note their analysis does not exclude some overlap between ICM/TE acquisition and EPI/PrE specification at E5. · stated by the authors
- The fate of the intermediate late-ICM cells, which have mixed EPI/PrE markers and middling NANOG:PDGFRA ratios, is unresolved. Whether apoptosis removes them or they persist before committing to a lineage was not tested. · stated by the authors
- Variable-gene selection was affected mostly by technical factors (library protocol, coverage, read distribution) where studies differed, and PC1 of the PSC-embryo PCA mainly reflected single-cell versus bulk processing. Intersecting per-study variable genes was used to reduce this bias. · stated by the authors
- Selecting only POU5F1-high clusters slightly underrepresented PrE cells, consistent with lower POU5F1 in PrE than in EPI; the final late ICM selection combined POU5F1-high and -medium clusters, so recovered lineage proportions depend on this step. · stated by the authors
- ESRRB as an early human ICM marker rests on unpublished non-human primate data (in preparation). The reclassified lineages are checked computationally against literature markers, not by new imaging of the same embryos. · noted in this summary
- PSC similarity scores compare bulk cultures with averaged single-cell embryo profiles without explicit batch correction. Absolute correlations and identity fractions may therefore partly reflect platform differences. · noted in this summary
Who should read:Read this if you use human embryo single-cell data as a reference for lineage annotation, or need to benchmark naïve versus conventional human pluripotent stem cell cultures against the preimplantation epiblast.
Reproducibility ledger
Nobody has reported trying to reproduce this paper yet. Did you replicate an experiment or run the code? Your report saves the next lab a week.
Sign in to file a reproducibility report.
Related papers
- Spatio-temporal analysis of human preimplantation development reveals dynamics of epiblast and trophectoderm
D. Meistermann, S. Loubersac, … L. David·bioRxiv·Apr 10, 2019
- A spatially resolved single cell atlas of human gastrulation
Richard C. V. Tyser, Elmir Mahammadov, … Shankar Srinivas·NaturePeer-reviewed·Nov 17, 2021
- The selection arena in early human blastocysts resolves the pluripotent inner cell mass
Manvendra Singh, Thomas J. Widmann, … Zsuzsanna Izsvák·bioRxiv·Feb 4, 2019
- Single cell analysis of lincRNA expression during human blastocyst differentiation identifies TERT (+) multi-lineage precursor cells
Jens Durruthy-Durruthy, Mark Wossidlo, … Gennadi V. Glinsky·bioRxiv·Oct 28, 2016
- A single-cell transcriptome atlas of human early embryogenesis
Yichi Xu, Tengjiao Zhang, … Weiyang Shi·Nature Cell BiologyPeer-reviewed·Mar 16, 2023
- Single-cell transcriptome analysis of human, marmoset and mouse embryos reveals common and divergent features of preimplantation development
Thorsten Boroviak, Giuliano Giuseppe Stirparo, … Paul Bertone·DevelopmentPeer-reviewed·Nov 1, 2018