CoChat
Sign Up
Cited by 4,490Single-cell and spatial transcriptomicsDevelopmental Biology and Gene RegulationGenomics and Chromatin Dynamics

The DOI matches its Crossref record; 68 of the 69 references it cites were found in a registry or an index, and 1 are not checked yet; none of the 31 sources its claims cite is checked for support yet.

The single-cell transcriptional landscape of mammalian organogenesis

Open access
Cao, Junyue, Spielmann, Malte, Qiu, Xiaojie, Huang, Xingfan, Ibrahim, Daniel M., Hill, Andrew J., Zhang, Fan, Mundlos, Stefan, Christiansen, Lena, Steemers, Frank J., Trapnell, Cole, Shendure, Jay·2019·Nature
DOI 10.1038/s41586-019-0969-xPMC PMC6434952
Open this paper in CoChat — free account

Abstract

Mammalian organogenesis is a remarkable process. Within a short timeframe, the cells of the three germ layers transform into an embryo that includes most of the major internal and external organs. Here we investigate the transcriptional dynamics of mouse organogenesis at single-cell resolution. Using single-cell combinatorial indexing, we profiled the transcriptomes of around 2 million cells derived from 61 embryos staged between 9.5 and 13.5 days of gestation, in a single experiment. The resulting ‘mouse organogenesis cell atlas’ (MOCA) provides a global view of developmental processes during this critical window. We use Monocle 3 to identify hundreds of cell types and 56 trajectories, many of which are detected only because of the depth of cellular coverage, and collectively define thousands of corresponding marker genes. We explore the dynamics of gene expression within cell types and trajectories over time, including focused analyses of the apical ectodermal ridge, limb mesenchyme and skeletal muscle. Data from single-cell combinatorial-indexing RNA-sequencing analysis of 2 million cells from mouse embryos between embryonic days 9.5 and 13.5 are compiled in a cell atlas of mouse organogenesis, which provides a global view of developmental processes occurring during this critical period.

Checked 2026-10-07 · 1 reference not checked yet

Overview

AI snapshot

Written from the paper’s full text.

TL;DR

This paper uses sci-RNA-seq3 to profile ~2 million cells from 61 mouse embryos staged E9.5–E13.5, producing the mouse organogenesis cell atlas (MOCA) of mid-gestational development. It identifies 38 major cell types, 572 non-doublet subtypes, and 56 developmental subtrajectories, with focused analyses of the apical ectodermal ridge and skeletal myogenesis.

Key findings

  • From one experiment, the authors recovered 2,058,652 mouse embryo cells plus 13,359 HEK293T/NIH/3T3 control cells, with a median of 35,272 cells per embryo; sequencing was shallow (~5,000 raw reads per cell) but yielded a median of 671 UMIs (519 genes) per cell.
  • MOCA contains 38 major cell types, 655 subclusters (572 non-doublet subtypes), and 56 developmental subtrajectories; 17,789 of 26,183 genes were differentially expressed across major cell types, including 2,863 cell type-specific marker genes.
  • Monocle 3 organized 1,524,792 high-quality cells (UMI > 400) into ten major trajectories, and the study matched 96 MCA fetal cell types to 58 MOCA subtypes and 48 BCA cell types to 68 MOCA subtypes.
  • Focused analyses identified AER subtype 6.23 (1,237 cells; 0.06% of MOCA) with novel markers Fndc3a, Adamts3, Slc16a10, Snap91, and Pou6f2, and a myogenic trajectory with 2,908 trajectory-dependent genes grouped into 14 patterns.

Methods

  • Nuclei were extracted from 61 C57BL/6 mouse embryos at E9.5, E10.5, E11.5, E12.5, or E13.5, fixed, and processed with sci-RNA-seq3 combinatorial indexing; the library was sequenced in one Illumina NovaSeq run yielding 11 billion reads.
  • Cells with UMI ≥ 200 were analyzed by Louvain clustering and t-SNE, followed by iterative subclustering; Scrublet was used for doublet detection, and marker genes were identified by differential expression.
  • Trajectory inference used Monocle 3 with UMAP, PAGA/Louvain partitioning, principal graph learning, pseudotime, and Moran’s I; whole-mount in situ hybridization validated selected markers.

Limitations

  • Paper-stated: only ~7% of cells entering the experiment were ultimately profiled, largely due to filtration steps; the matrix is sparse, the study does not exceed 1-fold coverage at any timepoint, and extremely rare cell types may be missed.
  • Paper-stated: annotations are preliminary because mid-gestational mouse development has not been extensively studied at single-cell resolution, many published markers have limited specificity, and whole disaggregated embryos make anatomical assignment challenging.
  • Evident from the text: sequencing was shallow (~5,000 raw reads per cell; median 671 UMIs), nuclear profiles were primarily intronic/nascent (59% intronic, 25% exonic), the estimated doublet rate was 10.3%, 13% of subtypes were annotated as likely artifacts, and trajectory discontinuity likely reflects missing ancestral states before E9.5.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1038/s41586-019-0969-x

Cites

68 of 69 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Verification

Claims the paper makes

0 of 31 cited sources checked · 0 supported · 0 flagged

  • Findingown data

    Despite shallow sequencing, the authors recovered a median of 671 UMIs and 519 genes per cell.

    “Despite shallow sequencing (~5,000 raw reads per cell; 46% duplicate rate), we recovered a median of 671 UMIs (519 genes) per cell”
    Quote in the paperNot a substantial claim
  • Findingown data

    13% of subtypes were annotated as likely artifacts because more than 10% of their cells were predicted doublets.

    “We annotated 13% of subtypes as likely artifacts (>10% of cells in these subtypes are predicted doublets”
    Quote in the paperNot a substantial claim
  • Findingown data

    At E9.5, cells corresponding to the primitive erythroid lineage from the yolk sac (cluster 26, marked by Hbb-bh1) were detected.

    “at E9.5, we detect cells corresponding to the primitive erythroid lineage, originating from the yolk sack (cluster 26; marked by Hbb−bh1).”
    Quote in the paperSubstantial claim
  • Findingown data

    Deeper sequencing of a subset of wells nearly doubled complexity to a median of 1,142 UMIs per cell.

    “3.7-fold deeper sequencing of a subset of wells nearly doubled complexity (to a median of 1,142 UMIs per cell; 87% duplicate rate).”
    Quote in the paperNot a substantial claim
  • Findingown data

    For the remaining 572 subtypes, the authors identified a median of 20 subtype-specific markers.

    “For the remaining 572 subtypes, we identified a median of 20 subtype-specific markers”
    Quote in the paperNot a substantial claim
  • It may state two separate claims as one. 43
  • Probably not a substantial claim. 30
  • The quote only partly states this claim. 1

  • It may state two separate claims as one.

    Findingown data

    Most subtypes can be distinguished from all other non-doublet subtypes based on marker gene sets and >4-fold expression differences, with 63% distinguished by 2 markers and 95% by 4 markers.

    “most subtypes can be distinguished from all 571 other non-doublet subtypes based on marker gene sets and >4-fold expression differences (63% with 2 markers, 95% with 4 markers”
    Quote in the paperSubstantial claim

  • It may state two separate claims as one.

    Findingown data

    Estimated shotgun cellular coverage of the mouse embryo was 0.8x at E9.5, 0.3x at E10.5, 0.2x at E11.5, 0.08x at E12.5, and 0.03x at E13.5.

    “our ‘shotgun cellular coverage’ of the mouse embryo is 0.8× at E9.5 (200K cells/embryo; 152K profiled across all replicates), 0.3× at E10.5 (1.1M cells/embryo; 378K profiled), 0.2× at E11.5 (2.6M cells/embryo; 616K profiled), 0.08× at E12.5 (6M cells/embryo; 475K profiled), and 0.03× at E13.5 (13M cells/embryo; 437K profiled).”
    Quote in the paperNot a substantial claim

  • It may state two separate claims as one.

    Findingown data

    Scrublet detected 4.3% likely doublet cells, corresponding to a doublet estimate of 10.3% including within- and between-cluster doublets.

    “We used Scrublet22 to detect 4.3% likely doublet cells, corresponding to a doublet estimate of 10.3% including both within- and between-cluster doublets”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [22] Wolock, S. L., Lopez, R. & Klein, A. M. Scrublet: computational identification of cell doublets in single-cell transcriptomic data. Preprint at https://www.biorxiv.org/content/10.1101/357368v1 (2018). source

  • Probably not a substantial claim. The paper’s conclusions may not depend on it.

    Findingown data

    Embryos were identified as male (n=31) or female (n=30).

    “Embryos were readily identified as male (n = 31) or female (n = 30)”
    Quote in the paperNot a substantial claim

  • It may state two separate claims as one.

    Findingown data

    From one experiment, the authors recovered 2,058,652 mouse embryo cells and 13,359 HEK293T or NIH/3T3 cells with UMI count at least 200.

    “From one experiment, we recovered 2,058,652 cells from mouse embryos and 13,359 cells from HEK293T or NIH/3T3 cells (UMI (unique molecular identifier) count ≥ 200).”
    Quote in the paperSubstantial claim

  • It may state two separate claims as one.

    Findingown data

    Pseudotemporal ordering of AER cells yielded a simple early-to-late trajectory and 710 differentially expressed genes.

    “We performed pseudotemporal ordering of AER cells, yielding a simple early-to-late trajectory and 710 differentially expressed genes”
    Quote in the paperNot a substantial claim

  • It may state two separate claims as one.

    Findingown data

    AER cells were detected at all timepoints and nearly all embryos.

    “Although detected at all timepoints and nearly all embryos, the estimated number of AER cells per embryo peaks between E10.5 and E11.5”
    Quote in the paperNot a substantial claim

  • It may state two separate claims as one.

    Findingown data

    The authors identified 2,863 cell type-specific marker genes, with a >5-fold cutoff yielding 932 marker genes.

    “Amongst these, we identified 2,863 cell type-specific marker genes (mean 75; those with >2-fold expression difference between first and second ranked cell type; a cutoff of >5-fold yielded 932 marker genes”
    Quote in the paperSubstantial claim

  • It may state two separate claims as one.

    Findingown data

    Because RNA was profiled in nuclei, 59% of UMIs per cell mapped to introns and 25% to exons.

    “As we are profiling RNA in nuclei, 59% of UMIs per cell strand-specifically mapped to introns, and 25% to exons.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim. The paper’s conclusions may not depend on it.

    Findingown data

    Later stage embryos exhibited somewhat reduced UMI counts.

    “Later stage embryos exhibited somewhat reduced UMI counts, possibly reflecting decreasing nuclear mRNA content”
    Quote in the paperNot a substantial claim

  • It may state two separate claims as one.

    Findingown data

    Cells from replicate embryos of the same developmental stage were similarly distributed, whereas cells from different stages were not.

    “Reassuringly, cells from replicate embryos of the same developmental stage were similarly distributed, whereas cells from different stages were not”
    Quote in the paperSubstantial claim

  • It may state two separate claims as one.

    Findingown data

    The 655 subtypes had a median of 1,869 cells and ranged from 51 to 65,894 cells.

    “The 655 subtypes consist of a median of 1,869 cells, and range from 51 (a subtype of notochord) to 65,894 (a subtype of connective tissue progenitors) cells”
    Quote in the paperNot a substantial claim

  • It may state two separate claims as one.

    Findingown data

    The 38 major cell types had a median of 47,073 cells, with the largest containing 144,648 cells and the smallest 1,000 cells.

    “The 38 major cell types are represented by a median of 47,073 cells, the largest containing 144,648 cells (connective tissue progenitors), and the smallest only 1,000 cells (neutrophils).”
    Quote in the paperNot a substantial claim

  • It may state two separate claims as one.

    Findingown data

    Most major cell types proliferated exponentially, while a few were transient and disappeared by E13.5.

    “While most major cell types proliferated exponentially, a few were transient and disappeared by E13.5”
    Quote in the paperNot a substantial claim

  • It may state two separate claims as one.

    Findingown data

    The primordial germ cell subcluster 16.13 contains 88 of 2,058,652 cells in the dataset and receives cells from all five timepoints.

    “For example, the primordial germ cell subcluster 16.13, which includes just 88 of 2,058,652 cells in the dataset, is contributed to by cells from all five timepoints.”
    Quote in the paperSubstantial claim

  • It may state two separate claims as one.

    Findingown data

    There are three neural crest trajectories, a hematopoietic trajectory, and four trajectories each corresponding to a single major cell type.

    “There are three neural crest trajectories, corresponding to sensory neurons, Schwann cell precursors and melanocytes. The hematopoietic trajectory includes megakaryocytes, erythrocytes and white blood cells, while the remaining four trajectories (endothelial, epithelial, hepatic, lens) each correspond to a single major cell type”
    Quote in the paperSubstantial claim

  • It may state two separate claims as one.

    Findingown data

    Primordial germ cells were readily identifiable despite their rarity, with subtypes 16.13 and 6.27 summing to 269 of 2,058,652 cells (0.01% of MOCA).

    “consider primordial germ cells, which were readily identifiable despite their rarity (subtypes 16.13 and 6.27, which sum to 269/2,058,652 cells or 0.01% of MOCA).”
    Quote in the paperSubstantial claim

  • It may state two separate claims as one.

    Findingown data

    UMAP processed a two-million-cell dataset in about 3 CPU hours, while t-SNE took more than 64 CPU hours.

    “We found that UMAP finished processing two million cells dataset in around 3 CPU hours while t-SNE took more than 64 CPU hours.”
    Quote in the paperNot a substantial claim

  • It may state two separate claims as one.

    Findingown data

    Some subtrajectories are single linear trajectories, many are branching, and some cell types appear to be generated via multiple parallel paths.

    “In some cases, we observe a single, simple linear trajectory. However, we also observe many examples of branching trajectories, as well as of cell types that appear to be generated via multiple parallel paths.”
    Quote in the paperNot a substantial claim

  • It may state two separate claims as one.

    Findingown data

    In the whole dataset, 4.3% of cells were detected as potential doublets, corresponding to an overall estimated doublet rate of 10.3% including within- and between-cluster doublets.

    “We detected 4.3% potential doublet cells in the whole data set, which corresponds to an overall estimated doublet rate of 10.3% (including both within- and between-cluster doublets).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [63] Mao, Q., Yang, L., Wang, L., Goodison, S. & Sun, Y. SimplePPT: a simple principal tree algorithm. In Proc. 2015 SIAM International Conference on Data Mining (eds Venkatasubramanian, S. & Ye, J.) 792–800 (SIAM, 2015). source

  • It may state two separate claims as one.

    Findingown data

    Cells expressing Myf5 co-localized with a subset of Pax7+ cells, consistent with the role of Myf5 in embryonic myogenesis.

    “Cells expressing Myf5 co-localized with a subset of Pax7+ cells, consistent with the role of Myf5 in embryonic myogenesis46.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [46] Comai, G., Sambasivan, R., Gopalakrishnan, S. & Tajbakhsh, S. Variations in the efficiency of lineage marking and ablation confound distinctions between myogenic cell populations. Dev. Cell 31, 654–667 (2014). source

  • Probably not a substantial claim. The paper’s conclusions may not depend on it.

    Findingown data

    After removing 12% of cells corresponding to doublet-annotated cells and/or subclusters, the authors iteratively reanalyzed the ten major trajectories.

    “After removing 12% of cells corresponding to doublet-annotated cells and/or subclusters, we iteratively reanalyzed the ten major trajectories”
    Quote in the paperNot a substantial claim

  • It may state two separate claims as one.

    Findingown data

    The estimated number of cells per embryo in each trajectory increases exponentially, and their proportions remain relatively stable except hepatocytes, which increase from 0.3% at E9.5 to 2.8% at E13.5.

    “Although the estimated number of cells per embryo in each trajectory increases exponentially, their proportions remain relatively stable, with the exception of hepatocytes, which markedly increase their contribution from 0.3% at E9.5 to 2.8% at E13.5”
    Quote in the paperNot a substantial claim

  • It may state two separate claims as one.

    Findingown data

    A myogenesis-specific Monocle 3 trajectory featured multiple focal concentrations of E9.5 cells, with later-stage cells distributed over several radiating paths.

    “we used Monocle 3 to construct a myogenesis-specific trajectory, which featured multiple focal concentrations of E9.5 cells, with cells from later stages distributed over several paths radiating outward”
    Quote in the paperNot a substantial claim

  • It may state two separate claims as one.

    Findingown data

    MOCA identifies over 500 subtypes and 56 developmental subtrajectories, each distinguished by multiple marker genes and collectively spanning essentially every organ system.

    “In the resulting atlas (MOCA), we identify over 500 subtypes of cells and 56 developmental subtrajectories, each distinguished by multiple marker genes and collectively spanning essentially every organ system.”
    Quote in the paperSubstantial claim

  • It may state two separate claims as one.

    Findingown data

    In a subset of wells, applying UMI error correction at edit distance 1 retained 99.4% of reads and changed per-cell UMI estimates by less than 1%.

    “Compared with skipping UMI error correction, 99.4% of reads remain after UMI error correction, which indicated to us that the error correction step only has minor impact on the estimated UMI counts per cell (less than 1%).”
    Quote in the paperSubstantial claim

  • It may state two separate claims as one.

    Findingown data

    Inhibitory neurons are connected to early CNS precursors by a bridge of neural progenitor cells, while radial glia project toward mature oligodendrocytes.

    “cell types found at later developmental timepoints such as inhibitory neurons are connected to early CNS precursors (radial glia) by a ‘bridge’ of neural progenitor cells; however, the same radial glial cells project in a different direction towards increasingly mature oligodendrocytes”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim. The paper’s conclusions may not depend on it.

    Findingown data

    The estimated nuclei extraction efficiency ranged from 60% to 85%.

    “The estimated nuclei extraction efficiency ranged from 60% to 85%.”
    Quote in the paperNot a substantial claim

  • It may state two separate claims as one.

    Findingown data

    Globally, 2,908 genes were expressed in a trajectory-dependent manner and grouped into 14 distinct patterns.

    “Globally, we detected 2,908 genes expressed in a trajectory-dependent manner (FDR < 0.05 and Moran’s I > 0.01) that grouped into 14 distinct patterns”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.

    Findingown data

    Among epithelial subtypes, 2 of 29 were doublet-derived; among endothelial subtypes, 5 of 16 were doublet-derived.

    “Doublet-derived subclusters (2/29 epithelial subtypes and 5/16 endothelial subtypes) are excluded from these plots”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim. The paper’s conclusions may not depend on it.

    Findingown data

    In the myogenic trajectory, 2,908 differentially expressed genes were clustered into 14 modules.

    “A total of 2,908 genes were clustered via hierarchical clustering. The dendrogram was cut into 14 modules using the cutree function in R”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim. The paper’s conclusions may not depend on it.

    Findingown data

    Some cells from complex trajectories have outlier pseudotime values more than 3 standard deviations above the mean.

    “Some cells from complex trajectories (mesenchymal trajectory, neural tube/notochord trajectory, epithelial trajectory and endothelial trajectory) show outlier pseudotime values (more than 3 standard deviation higher than the mean values).”
    Quote in the paperNot a substantial claim

  • It may state two separate claims as one.

    Findingown data

    With 23,207 reads per cell, the median UMI count was 5,461 for HEK293T cells and 5,087 for NIH/3T3 cells.

    “At a sequencing depth of 23,207 reads per cell, we observed a median of 5,461 UMIs per HEK293T cell and 5,087 UMIs per NIH/3T3 cell”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim. The paper’s conclusions may not depend on it.

    Findingown data

    With downsampling, the number of clusters and subclusters identified using the same parameters fell from 38 to 27 and from 16 to 12, respectively.

    “The number of clusters and subclusters identified with the same parameters drops from 38 (a, bottom plot) to 27 (c) and 16 (b, bottom plot) to 12 (c), respectively.”
    Quote in the paperNot a substantial claim

  • It may state two separate claims as one.

    Findingown data

    Among 655 subclusters, 644 (98%) had a marker with 2-fold higher expression, and 441 (67%) had a marker with 5-fold higher expression.

    “644 of 655 sub-clusters (98%) have at least one such gene marker with a 2-fold difference, and 441 of 655 (67%) have at least one such marker with a 5-fold difference.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim. The paper’s conclusions may not depend on it.

    Findingown data

    The study identified 40 clusters.

    “40 clusters were identified.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim. The paper’s conclusions may not depend on it.

    Findingown data

    The detected doublet cell rate showed a modest positive correlation with the number of cells profiled per reverse-transcription well (Spearman's rho 0.35).

    “The detected doublet cell rate was modestly correlated with number of cells profiled per well during reverse transcription (Spearman’s rho: 0.35).”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim. The paper’s conclusions may not depend on it.

    Findingown data

    Merging two sensory-neuron PAGA groups and two blood-cell PAGA groups yielded ten supergroups.

    “We merged two groups corresponding to sensory neurons (12 & 3), and another two groups corresponding to blood cells (6 & 7), as each pair was closely located in UMAP space upon visual inspection, yielding the ten supergroups shown in a similar heatmap in Fig. 4b.”
    Quote in the paperNot a substantial claim

  • The quote only partly states this claim.

    Findingown data

    The proportions of cells per supergroup remained relatively stable from E9.5 to E13.5, except for hepatocytes.

    “Although the estimated number of cells per embryo in each of these supergroups increases exponentially, their proportions remain relatively stable”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim. The paper’s conclusions may not depend on it.

    Findingown data

    Two clusters appeared to correspond to the definitive erythroid lineage and were merged.

    “Another two clusters both appeared to correspond to the definitive erythroid lineage and were merged.”
    Quote in the paperNot a substantial claim

  • It may state two separate claims as one.

    Interpretationown data

    The nuclear profiles may primarily reflect nascent transcription, temporally offset but predictive of the cellular transcriptome.

    “Our profiles may thus primarily reflect nascent transcription, temporally offset but also predictive21 of the cellular transcriptome.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [21] La Manno, G. et al. RNA velocity of single cells. Nature 560, 494–498 (2018). source

  • It may state two separate claims as one.

    Interpretationreasoning

    MOCA subtypes that failed to match MCA cell types tended to derive from earlier stages or be rare, while MCA cell types that failed to match were mostly tissue-specific immune or epithelial cells, potentially because they emerge after E13.5.

    “As expected, MOCA subtypes that failed to match MCA cell types tended to derive from earlier stages (e.g. neural tube) or were rare (e.g. lens), while MCA cell types that failed MOCA subtypes were mostly tissue-specific immune or epithelial cells, potentially because they emerge after E13.5.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.

    Study design and conductown data

    Only about 7% of cells entering the experiment were ultimately profiled, with losses largely due to filtration steps intended to remove aggregates of nuclei.

    “A limitation is that only ~7% of cells entering the experiment were ultimately profiled, with losses largely consequent to filtration steps intended to remove aggregates of nuclei.”
    Quote in the paperNot a substantial claim

  • It may state two separate claims as one.

    Premisecitation

    Recently released mouse scRNA-seq atlases are mostly restricted to adult organs and do not attempt to characterize the emergence and dynamics of cell types during development.

    “However, although two scRNA-seq atlases of mouse were recently released10,11, they are mostly restricted to adult organs, and do not attempt to characterize the emergence and dynamics of cell types during development.”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites [10] Han, X. et al. Mapping the Mouse Cell Atlas by Microwell-seq. Cell 172, 1091–1107 (2018). source
    • cites [11] The Tabula Muris Consortium, Quake, S. R., Wyss-Coray, T. & Darmanis, S. Transcriptomic characterization of 20 organs and tissues from mouse at single cell resolution creates a Tabula Muris. Preprint at https://www.biorxiv.org/content/10.1101/237446v2 (2018).

  • It may state two separate claims as one.

    Interpretationown data

    The MCA and MOCA atlases inform one another: MCA anatomical resolution helps localize MOCA subtypes, and MOCA developmental focus informs the embryonic origin of MCA cell types.

    “Nonetheless, the atlases unquestionably inform one another, as the MCA’s anatomical resolution is useful for localizing MOCA subtypes, while MOCA’s developmental focus informs the embryonic origin of MCA cell types”
    Quote in the paperSubstantial claim

  • It may state two separate claims as one.

    Interpretationown data

    An additional path traversed by E9.5 cells expressed Lhx2, Tbx1, and Pitx2 but low Pax3, feeds upstream of the Myf5 and Myod1 segments, and possibly corresponds to pharyngeal mesoderm.

    “An additional path traversed by cells from E9.5, which expressed Lhx2, Tbx1, and Pitx2 but very low levels of Pax3, feeds into the trajectory just upstream of the Myf5 and Myod1 segments, and possibly corresponds to pharyngeal mesoderm45.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [45] Braun, T. & Gautel, M. Transcriptional mechanisms regulating skeletal muscle differentiation, growth and homeostasis. Nat. Rev. Mol. Cell Biol. 12, 349–361 (2011). source

  • Probably not a substantial claim. The paper’s conclusions may not depend on it.

    Study design and conductnone given

    No statistical methods were used to predetermine sample size.

    “No statistical methods were used to predetermine sample size.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.

    Premisecitation

    UMAP has O(N) algorithmic complexity, while t-SNE has O(Nlog(N)) complexity.

    “The UMAP algorithm itself is also more efficient (the algorithm complexity of UMAP is O(N) vs. O(Nlog(N)) for t-SNE).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [66]

  • Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

    Study design and conductnone given

    Monocle 3's partitionCells() implementation scales to millions of cells.

    “Monocle 3’s implementation of the above procedures (in the partitionCells() function) scales to millions of cells.”
    Quote in the paperNot a substantial claim

  • It may state two separate claims as one.

    Premisenone given

    Most contemporary trajectory reconstruction algorithms assume a continuous manifold and do not allow convergence of cell fates.

    “Most contemporary algorithms for trajectory reconstruction assume a continuous manifold (whereas our data begin at E9.5, and therefore are missing at least some ancestral states) and do not allow for convergence of cell fates”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.

    Study design and conductnone given

    Investigators were blinded to group allocation during data collection and analysis, with embryo collection and sci-RNA-seq3 analysis performed by two different researchers.

    “Investigators were blinded to group allocation during data collection and analysis: embryo collection and sci-RNA-seq3 analysis were performed by two different researchers.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim. The paper’s conclusions may not depend on it.

    Study design and conductnone given

    Embryos were collected from at least three independent litters per developmental stage.

    “Embryos were collected from at least three independent litters per development stage.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim. The paper’s conclusions may not depend on it.

    Study design and conductnone given

    Embryos used in the experiment were randomized before sample preparation.

    “Embryos used in experiment were randomized before sample preparation.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim. The paper may give no reason why it holds.

    Premisenone given

    Mid-gestational mouse development has not previously been extensively studied at single cell resolution.

    “Mid-gestational mouse development has not previously been extensively studied at single cell resolution”
    Quote in the paperNot a substantial claim

  • It may state two separate claims as one.

    Interpretationown data

    The AER subtrajectory projects out of surface ectoderm and back into epidermis, consistent with its transitory nature.

    “Notably, the AER subtrajectory projects out of surface ectoderm and then back into epidermis, consistent with its transitory nature.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim. The paper’s conclusions may not depend on it.

    Study design and conductreasoning

    Monocle 3 can learn principal graphs with loops instead of requiring the trajectory to be a tree.

    “The third major enhancement is that Monocle 3 can learn principal graphs with loops instead of requiring that the trajectory be a tree.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

    Premisenone given

    Many published markers have limited specificity.

    “many published markers have limited specificity.”
    Quote in the paperNot a substantial claim

  • It may state two separate claims as one.

    Interpretationown data

    CNS excitatory and inhibitory neurons appear to develop through multiple convergent trajectories, possibly due to maturation in multiple anatomical locations.

    “both CNS excitatory and inhibitory neurons appear to develop through multiple, convergent trajectories, possibly due to the maturation in multiple anatomical locations.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim. The paper’s conclusions may not depend on it.

    Study design and conductreasoning

    The authors' annotations should be regarded as preliminary.

    “Second, although we are reasonably confident in our annotations, they should be regarded as preliminary.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.

    Premisecitation

    UMAP preserves longer-range distance relationships, whereas t-SNE mainly places highly similar cells in the same regions of a low-dimensional space.

    “However, where t-SNE mainly aims to place highly similar cells in the same regions of a low-dimensional space, UMAP also preserves longer-range distance relationships.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [66]

  • Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.

    Premisenone given

    Running time and trajectory detail depend on the number of principal graph nodes; more nodes generally yield a more accurate tree but increase running time.

    “Running time and fine detail in the trajectory will depend on the number principal graph nodes; more nodes generally results in a more accurate tree but at increased running time.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim. The paper’s conclusions may not depend on it.

    Premisecitation

    UMAP's visualization quality is competitive with t-SNE.

    “Its visualization quality is competitive with the popular t-SNE (t-stochastic neighbor embedding) method used widely in single-cell transcriptomics.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [66]

  • Probably not a substantial claim. The paper’s conclusions may not depend on it.

    Study design and conductown data

    Embryos were readily separated into females and males based on reads mapping to Xist versus chrY genes.

    “Embryos were readily separated into females (more reads mapping to Xist than chrY genes) and males (more reads mapping to chrY genes than Xist).”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.

    Premisecitation

    PAGA constructs a graph whose vertices are Louvain communities, linking two communities when their cells are neighbors in the kNN graph more often than expected under a simple binomial model.

    “PAGA then constructs a graph in which the vertices are Louvain communities. Two vertices are linked with an edge in the PAGA graph when the cells in the respective communities are neighbors in the kNN graph more frequently than would be expected under a simple binomial model68.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [68]

  • Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

    Study design and conductnone given

    The projection strategy developed for cell-wise pseudotime is applicable to datasets with millions of cells.

    “In order to calculate cell-wise pseudotime, we developed a projection strategy which is applicable to datasets with millions of cells.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim. It may be too vague to check against evidence. The paper’s conclusions may not depend on it. It may state two separate claims as one.

    Study design and conductown data

    The simple heuristic pruning procedure is effective and easier for users to understand how to control.

    “Although SimplePPT does provide tuning parameters that control graph smoothness to a certain extent, we have found that a simple heuristic pruning procedure is effective and easier for users to understand how to control.”
    Quote in the paperNot a substantial claim
66 of 66

  • The authors state this as an assumption.

    About 100 cells are required to detect a cell type.

    “Assuming that ~100 cells are required to detect a cell type”

  • The authors state this as an assumption.

    The cell type in question is present at only one timepoint.

    “and that the cell type in question is only present at one timepoint”

  • The authors state this as an assumption.

    Gene expression follows a negative binomial distribution in the AER spline fitting.

    “assuming gene expression following the negative binomial distribution”
3 of 3

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

  • Monocle3 pseudotime applied to 2,190 radial glia lineage cells identified three trajectories corresponding to distinct cell fates, termed the ependymal (EPD) trajectory, astrocyte (ASC) trajectory, and neural intermediate progenitor cell (NPC) trajectory.

    “we applied Monocle336 pseudotime to 2,190 radial glia lineage cells (Figure 4B). We identified three trajectories corresponding to distinct cell fates, termed the ependymal (EPD) trajectory, astrocyte (ASC) trajectory, and the neural intermediate progenitor cell (NPC) trajectory.”

    From hdWGCNA identifies co-expression networks in high-dimensional transcriptomics data (2023) · cites this paper as [36]

  • Recent cell atlases using high-throughput scRNA-seq and scATAC-seq have provided an increasingly detailed view of mouse developmental gene expression and gene regulatory dynamics.

    “Recent cell atlases using high-throughput single-cell RNA-sequencing (scRNA-seq) and single-cell assay for transposase accessible chromatin sequencing (scATAC-seq) technologies have provided an increasingly detailed view of mouse developmental gene expression and gene regulatory dynamics”

    From Spatiotemporal transcriptomic atlas of mouse organogenesis using DNA nanoball-patterned arrays (2022) · cites this paper as Cao et al., 2019

  • The stages E9.5-E16.5 include most of the key events in mouse organogenesis.

    “These stages include most of the key events in mouse organogenesis (Cao et al., 2019).”

    From Spatiotemporal transcriptomic atlas of mouse organogenesis using DNA nanoball-patterned arrays (2022) · cites this paper as Cao et al., 2019

  • Spatial pseudotime analysis of progenitor cells at the three stages using monocle 3 confirmed the results.

    “Spatial pseudotime analysis of progenitor cells in these three stages using monocle 3 (Cao et al., 2019) confirmed these results (Figures 5D-5F and S7E).”

    From Spatiotemporal transcriptomic atlas of mouse organogenesis using DNA nanoball-patterned arrays (2022) · cites this paper as Cao et al., 2019

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

  • Ligation is described as completed before its reaction is set up

    Minor

    The protocol says the ligation reaction has finished before describing redistribution into wells containing ligase and stating that the ligation reaction was then performed.

  • Endothelial example contains more cells than the reported endothelial total

    Minor

    The caption calls the 74,651-cell example endothelial subclusters, but elsewhere reports only 35,878 endothelial cells in total.

2 of 2

Citations

Cites

Flagged first, then by citations

Cited by

Sorted by citations

Related

Metadata

Authors
Cao, J., Spielmann, M., Qiu, X., Huang, X., Ibrahim, D. M., Hill, A. J., Zhang, F., Mundlos, S., Christiansen, L., Steemers, F. J., Trapnell, C., & Shendure, J.
Affiliations
  • University of Washington (US)
  • Max Planck Institute for Molecular Genetics (DE)
  • Illumina (United States) (US)
  • Howard Hughes Medical Institute (US)
Corresponding author
  • Trapnell, Cole
  • Shendure, Jay
ORCID
Venue
Nature
Publisher
Springer Science and Business Media LLC
Published
2019-02-20
Volume
566
Issue
7745
Pages
496-502
ISSN
  • 0028-0836 (ISSN-L)
  • 1476-4687
DOI
10.1038/s41586-019-0969-x
PMC
PMC6434952
OpenAlex
W2916020270
Semantic Scholar
cc5f35ade6b0ecee863956c541fb239c1db1d593
Cited by
4,490
References
64
Citations by year
2026: 797 · 2025: 901 · 2024: 809 · 2023: 645 · 2022: 568 · 2021: 421 · 2020: 267 · 2019: 76
FWCI
111.88
Type
journal article
Language
en
Open access
Licence
Locations
  • Nature (publishedVersion)
  • PubMed (publishedVersion)
  • Europe PMC (PubMed Central) (acceptedVersion)
  • PubMed Central (acceptedVersion)
  • RePEc: Research Papers in Economics (submittedVersion)
Funders
  • Howard Hughes Medical Institute
  • National Heart, Lung, and Blood Institute (R01HL118342, U54HL145611)
  • National Human Genome Research Institute (DP1HG007811, R01HG006283)
  • National Institute of Diabetes and Digestive and Kidney Diseases (U54DK107979)
  • Eunice Kennedy Shriver National Institute of Child Health and Human Development (DP2HD088158)
Topics
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Single-cell and spatial transcriptomics
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Developmental Biology and Gene Regulation
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Genomics and Chromatin Dynamics
Keywords
mouse embryo, single-cell transcriptomics, cell atlas, developmental trajectories, cell types, marker genes, apical ectodermal ridge, limb mesenchyme, skeletal muscle development, gene expression dynamics, three germ layers
MeSH terms
Animals, Ectoderm, Embryo, Mammalian, Female, Genetic Markers, Male, Mesoderm, Organ Specificity, Time Factors, Sequence Analysis, RNA, Muscle, Skeletal, Gene Expression Regulation, Developmental, Muscle Development, Organogenesis, Mice, Single-Cell Analysis, Transcriptome

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

A language model drew the claims from the paper on 2026-10-07, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-10-07 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review