Dissecting cell identity via network inference and in silico gene perturbation
Open accessThe DOI matches its Crossref record; 91 of the 91 references it cites were found in a registry or an index; 27 of the 108 sources its claims cite were checked for support.
Abstract
Cell identity is governed by the complex regulation of gene expression, represented as gene-regulatory networks 1 . Here we use gene-regulatory networks inferred from single-cell multi-omics data to perform in silico transcription factor perturbations, simulating the consequent changes in cell identity using only unperturbed wild-type data. We apply this machine-learning-based approach, CellOracle, to well-established paradigms—mouse and human haematopoiesis, and zebrafish embryogenesis—and we correctly model reported changes in phenotype that occur as a result of transcription factor perturbation. Through systematic in silico transcription factor perturbation in the developing zebrafish, we simulate and experimentally validate a previously unreported phenotype that results from the loss of noto , an established notochord regulator. Furthermore, we identify an axial mesoderm regulator, lhx1a . Together, these results show that CellOracle can be used to analyse the regulation of cell identity by transcription factors, and can provide mechanistic insights into development and differentiation.
Overview
CoChat Summary
TL;DR
CellOracle combines inferred gene-regulatory networks with in silico transcription-factor perturbation to simulate shifts in cell identity from unperturbed single-cell multi-omics data. The authors validate it in mouse and human haematopoiesis and zebrafish embryogenesis, then use it to predict and experimentally confirm a noto loss-of-function prechordal plate expansion and identify lhx1a as an axial mesoderm regulator.
Key findings
- CellOracle GRN inference benchmarked against 1,298 ChIP–seq datasets for 80 regulatory factors across 5 tissues achieved AUROC 0.66–0.85 with a promoter base GRN and 0.73–0.91 with an scATAC-seq base GRN.
- In a 2,730-cell mouse myeloid progenitor atlas, CellOracle reproduced known Spi1/Gata1 lineage-switching phenotypes; among the top 30 TFs ranked by systematic KO simulation, 85% were reported regulators of myeloid differentiation.
- In zebrafish, systematic KO simulation of 232 active TFs ranked noto top by degree centrality in notochord; noto LOF simulation predicted prechordal plate expansion, validated in flhn1/n1 mutants (notochord −98%, P = 5.55 × 10−52; prechordal plate +38.6%, P = 1.07 × 10−4) and noto crispants.
- lhx1a LOF simulation predicted inhibited early axial mesoderm differentiation; lhx1a crispants showed early notochord expansion (+70.2%, P = 1.34 × 10−35) and depletion of later notochord (−15.3%, P = 3.83 × 10−3) and prechordal plate (−24.7%, P = 1.28 × 10−7).
Methods
- CellOracle builds cluster-wise regularized linear GRN models using scRNA-seq plus scATAC-seq or promoter databases, propagates TF perturbation signals through the network, estimates transition probabilities, and projects simulated cell-identity shifts as 2D vectors; perturbation scores compare simulation vectors with pseudotime differentiation vectors.
- Validation used mouse and human haematopoiesis datasets, a 38,731-cell zebrafish embryogenesis atlas, a 38,606-cell wild-type zebrafish reference atlas, flhn1/n1 mutants, and CRISPR–Cas9 F0 crispants; experimental readouts included scRNA-seq, NMF, HCR, qRT-PCR and whole-mount in situ hybridization.
- Benchmarking compared CellOracle with WGCNA, DCOL, GENIE3 and SCENIC using ground-truth GRNs from ChIP–seq data; simulations also used randomized GRNs, no signal propagation, varying signal-propagation rounds, and different dimensionality-reduction methods.
Limitations
- Stated: CellOracle can only analyse cell states present in the input scRNA-seq data, and simulation is limited by input data availability and quality, such as unknown TF-binding motifs or sparse TF expression; the authors note foxa3 was filtered out owing to low expression.
- Stated: experimental validation is essential where possible; the authors stress that predictions should be validated experimentally.
- Evident: the simulation output is deliberately limited to a 2D vector rather than absolute expression changes; the negative perturbation score may oversimplify TF roles (e.g. Elf1), and CellOracle did not detect a reported erythrocyte progenitor depletion after Spi1 KO, probably owing to proliferation changes not predicted by the method.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Figures


















Verification
Claims the paper makes
14 Valid · 2 Not valid · 54 Not checked · 208 Own results and interpretations
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“To benchmark our GRN inference method, we generated a comprehensive transcriptional ground-truth GRN using 1,298 chromatin immunoprecipitation followed by sequencing (ChIP–seq) datasets for 80 regulatory factors across 5 different tissues14.”
§ Dissecting cell identity via network inference and in silico gene perturbation > GRN inference and benchmarking with CellOracle
What the sources say
- Partly supportedcites [14] Oki, S. et al. ChIP-Atlas: a data-mining suite powered by full integration of public ChIP-seq data. EMBO Rep. 19, e46255 (2018). sourceExperimental — accuracy not yet measured
Our copy of ChIP‑Atlas confirms it is a fully integrated public ChIP‑seq/DNase‑seq resource (76,217 SRXs across six organisms, Abstract; Overview of the data set and design of analyses, Table 1) with curated antigen and cell‑type metadata and a TR–gene ('Target Genes') integration that 'provides insight into gene regulatory networks' — so it can be the dataset source behind a ChIP‑seq‑derived regulatory network. However, no passage states the specific quantities the claim attributes here (1,298 datasets, 80 regulatory factors, 5 tissues), nor that the resource constitutes or was used to generate a 'ground‑truth' GRN; on the contrary, the Results section 'TR–gene and TR–gene interactions' cautions that 'the genes listed by "Target Genes" are not necessarily functional targets of a given TR and that actual regulation ... would need to be confirmed experimentally'. Searched: 1,298/80 factors/five tissues, ground-truth GRN benchmark, gold-standard TF–target set. (LLM, conf=0.65)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Partly supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“In this respect, gene-regulatory network (GRN) modelling approaches are promising as they reconstruct systematic gene–gene associations from unperturbed single-cell omics data6–11.”
§ Dissecting cell identity via network inference and in silico gene perturbation > Main
What the sources say
- Supportedcites [6] Fiers, M. W. E. J. et al. Mapping gene regulatory networks from single-cell omics data. Brief. Funct. Genomics 17, 246–254 (2018). source
The “GRN inference from scRNA-seq data” section describes reconstructing networks from pairwise associations and co-expression in single-cell data. The Introduction discusses perturbation assays separately as a way to validate predictions, consistent with the claim that these associations can be inferred from unperturbed data. (LLM, conf=0.90)
- Supportedcites [7] Aibar, S. et al. SCENIC: single-cell regulatory network inference and clustering. Nat. Methods 14, 1083–1086 (2017). source
The Abstract describes SCENIC as reconstructing gene regulatory networks from single-cell RNA-seq data and identifies co-expressed genes as its first step. The Online Methods (GENIE3; page unavailable) specify an expression matrix as the input for network inference, consistent with reconstructing gene associations without requiring perturbation data. The paper addresses single-cell transcriptomics specifically, rather than every type of single-cell omics data. (LLM, conf=0.86)
- Supportedcites [8] Iacono, G., Massoni-Badosa, R. & Heyn, H. Single-cell transcriptomics unveils gene regulatory network plasticity. Genome Biol. 20, 110 (2019). source
The Results section describes inferring regulatory networks from single-cell expression data by identifying gene-to-gene correlations and retaining significant correlations as network edges. The Background also states that the approach was applied to healthy tissue datasets, supporting the claim that it can reconstruct associations from unperturbed single-cell data. (LLM, conf=0.91)
- Supportedcites [9] Fleck, J. S. et al. Inferring and perturbing cell fate regulomes in human brain organoids. Nature https://doi.org/10.1038/s41586-022-05279-8 (2022). source
The ‘Regulatory network inference with Pando’ section describes using multimodal single-cell measurements and regression to infer relationships among transcription-factor expression, binding-site accessibility and target-gene expression. The Methods section (‘GRN inference with Pando’; page unavailable) says the resulting model coefficients define interactions in a regulatory graph. This supports the claim that GRN modelling reconstructs gene–gene regulatory associations from single-cell omics data without requiring perturbations for the inference step. (LLM, conf=0.90)
- Partly supportedcites [10] Kartha, V. K. et al. Functional inference of gene regulation using single-cell multi-omics. Cell Genom. 2, 100166 (2022). sourceExperimental — accuracy not yet measured
The Discussion (page unavailable) says FigR constructs gene-regulatory networks from correlations across single-cell chromatin and RNA profiles, supporting the claim about reconstructing gene associations. But the paper describes an atlas of immunological stimulation and analyses across stimulus conditions, so it does not establish the claim's qualification that the data are unperturbed. (LLM, conf=0.82)
- Supportedcites [11] González-Blas, C. B. et al. SCENIC+: single-cell multiomic inference of enhancers and gene regulatory networks. Preprint at bioRxiv https://doi.org/10.1101/2022.08.19.504505 (2022). source
The SCENIC+ methods (p. 2) describe inferring TF-to-gene relationships from single-cell gene-expression and chromatin-accessibility data using correlation and regression, then combining them into gene regulatory networks. The Discussion (p. 13) also describes GRN methods that search expression data for putative TF–target-gene relationships. This supports the claim as a broad description of network modelling from observational single-cell data. (LLM, conf=0.91)
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“To support GRN inference without requiring sample-specific scATAC-seq datasets, we have assembled a base GRN from a mouse scATAC-seq atlas13.”
§ Dissecting cell identity via network inference and in silico gene perturbation > GRN inference and benchmarking with CellOracle
What the sources say
- Supportedcites [13] Cusanovich, D. A. et al. A single-cell atlas of in vivo mammalian chromatin accessibility. Cell 174, 1309–1324 (2018). source
The Results describe profiling chromatin accessibility in single nuclei from 13 mouse tissues using sci-ATAC-seq, identifying the cited work as a mouse single-cell ATAC-seq atlas. The sentence attributes the assembly of the base GRN and its use without sample-specific datasets to the citing authors, not to this source. (LLM, conf=0.92)
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“Because of this second constraint, we cannot use pre-existing GRN inference algorithms, such as GENIE3 and GRNboost (refs. 7,51).”
§ Dissecting cell identity via network inference and in silico gene perturbation > Methods > Context-dependent GRN inference using scRNA-seq data
What the sources say
- Supportedcites [7] Aibar, S. et al. SCENIC: single-cell regulatory network inference and clustering. Nat. Methods 14, 1083–1086 (2017). source
The Online Methods (GRNboost) identify GRNboost and GENIE3 as gene-regulatory-network inference methods and explain that they use boosted regression trees and random forests, respectively, rather than linear regression. The citing authors’ inability to use them under their constraint is their own conclusion; the citation supports the identification of the algorithms and their methods. (LLM, conf=0.92)
- Supportedcites [51] Huynh-Thu, V. A., Irrthum, A., Wehenkel, L. & Geurts, P. Inferring regulatory networks from expression data using tree-based methods. PLoS ONE 5, e12776 (2010). source
The Introduction identifies GENIE3 as a GRN inference method using ensembles of regression trees and says it can model non-linear relationships. This supports citing the paper as an example of an algorithm unsuitable for the citing authors’ linear-regression constraint; that constraint is the citing authors’ own premise. The paper is not being cited here for GRNboost. (LLM, conf=0.91)
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
The sources with a verdict support this claim; the rest weren’t checked yet, have no full text we could read, or their verdicts aren’t rated yet.
In the paper
“Cebpa promotes GM lineage differentiation while inhibiting ME lineage differentiation16,72, and promoting granulocyte differentiation in particular15.”
§ Dissecting cell identity via network inference and in silico gene perturbation > Methods > Comparison between CellOracle haematopoiesis KO simulation results and previous reports
What the sources say
- Not checked yetcites [16] Paul, F. et al. Transcriptional heterogeneity and lineage commitment in myeloid progenitors. Cell 163, 1663–1677 (2015). source
- Supportedcites [72] Hyung, C. S. et al. C/EBPα determines hematopoietic cell fate in multipotential progenitor cells by inhibiting erythroid differentiation and inducing myeloid differentiation. Blood 107, 4308 (2006). source
In the section on lineage switching of EML cells, the paper reports that C/EBPα overexpression promoted more rapid and complete granulocyte differentiation than the comparison treatment. This directly supports the attributed claim. (LLM, conf=0.94)
- Not checked yetcites [15] Orkin, S. H. & Zon, L. I. Hematopoiesis: an evolving paradigm for stem cell biology. Cell 132, 631–644 (2008). source
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
147 more claims · 3 with their cited sources checked
Create a free account to see all 278 claims and to verify the 81 references still open.
- It may state two separate claims as one. 87
- Probably not a substantial claim. 53
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?
The source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“CellOracle surpassed other GRN-inference algorithms such as PIDC, GENIE3, and GRNBoost2 in accurately identifying biological ground truth [44]”
From A mini-review on perturbation modelling across single-cell omic modalities (2024) · cites this paper as [44]
Experimental — accuracy not yet measuredThe cited source says what the claim attributes to it.
In the citing paper
“By leveraging gene-regulatory networks derived from these data types, CellOracle performs in silico transcription factors perturbations, effectively simulating cell state changes in response to TF perturbations.”
From A mini-review on perturbation modelling across single-cell omic modalities (2024) · cites this paper as [44]
In the citing paper
“For instance, CellOracle can simulate cell identity transitions following digital (in silico) TF perturbations such as knockout or overexpression.”
From A mini-review on perturbation modelling across single-cell omic modalities (2024) · cites this paper as [44]
In the citing paper
“whereas in silico ChIP–seq predicts transcription factor binding at individual cis-regulatory elements179.”
From The technological landscape and applications of single-cell multi-omics (2023) · cites this paper as [179]
In the citing paper
“Similarly, CellOracle179 leverages multi-omics data, either collected in parallel or integrated computationally, to learn gene regulatory networks and infer the results of in silico perturbations, whereas in silico ChIP–seq predicts transcription factor binding at individual cis-regulatory elements179.”
From The technological landscape and applications of single-cell multi-omics (2023) · cites this paper as [179]
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.
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ Dissecting cell identity via network inference and in silico gene perturbation > Methods > Simulation of cell identity following perturbation of regulatory genes
“This process consists of four steps: (i) data preprocessing; (ii) signal propagation within the GRN; (iii) estimation of transition probabilities; and (iv) analysis of simulated transition in cell identity.”
§ Dissecting cell identity via network inference and in silico gene perturbation > Methods > Simulation of cell identity following perturbation of regulatory genes
“(v)Calculation of vector field”
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ Dissecting cell identity via network inference and in silico gene perturbation > Methods > Preparation of input data for GRN inference
“Cells were subsampled for each tissue on the basis of the original tissue-type annotation: spleen, lung, muscle, liver and kidney.”
§ Dissecting cell identity via network inference and in silico gene perturbation > Methods > Ground-truth data preparation for GRN benchmarking
“Here, we selected tissues commonly available in the Tabula Muris scRNA-seq dataset, mouse sci-ATAC-seq atlas data and ground-truth datasets: heart, kidney, liver, lung and spleen.”
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ Dissecting cell identity via network inference and in silico gene perturbation > Systematic TF KO simulations in zebrafish
“c, Degree centrality scores were used to rank the top 30 TFs in notochord (left) and prechordal plate (right).”
§ Dissecting cell identity via network inference and in silico gene perturbation > Systematic TF KO simulations in zebrafish
“Black text denotes TFs. Grey text denotes non-TFs.”
Citations
Cites
Comprehensive Integration of Single-Cell Data
Stuart, T., Butler, A., Hoffman, P., Hafemeister, C., Papalexi, E., Mauck, W. M., Hao, Y., Stoeckius, M., Smibert, P., & Satija, R. (2019) · Cell · doi:10.1016/j.cell.2019.05.031 · Cited by 15,453
SCANPY: large-scale single-cell gene expression data analysis
Wolf, F. A., Angerer, P., & Theis, F. J. (2018) · Genome Biology · doi:10.1186/s13059-017-1382-0 · Cited by 8,881
Scenic: single-cell regulatory network inference and clustering.
Aibar, S., González-Blas, C. B., Moerman, T., Huynh-Thu, V. A., Imrichova, H., Hulselmans, G., Rambow, F., Marine, J.-C., Geurts, P., Aerts, J., et al. (2017) · Nature Methods · doi:10.1038/nmeth.4463 · Cited by 6,520
A General Framework for Weighted Gene Co-Expression Network Analysis
Zhang, B., & Horvath, S. (2005) · Statistical Applications in Genetics and Molecular Biology · doi:10.2202/1544-6115.1128 · Cited by 5,684
Hafemeister, C., & Satija, R. (2019) · Genome Biology · doi:10.1186/s13059-019-1874-1 · Cited by 4,479
La Manno, G., Soldatov, R., Zeisel, A., Braun, E., Hochgerner, H., Petukhov, V., Lidschreiber, K., Kastriti, M. E., Lönnerberg, P., Furlan, A., et al. (2018) · Nature · doi:10.1038/s41586-018-0414-6 · Cited by 4,118
Jacomy, M., Venturini, T., Heymann, S., & Bastian, M. (2014) · PLoS ONE · doi:10.1371/journal.pone.0098679 · Cited by 2,909
Single-cell transcriptomics of 20 mouse organs creates a Tabula Muris
The Tabula Muris Consortium, coordination, O., coordination, L., processing, O. C. A., sequencing, L. P. A., analysis, C. D., annotation, C. T., Writing group, Supplemental text writing group, & Principal investigators (2018) · Nature · doi:10.1038/s41586-018-0590-4 · Cited by 2,842
High-resolution in situ hybridization to whole-mount zebrafish embryos
Thisse, C., & Thisse, B. (2008) · Nature Protocols · doi:10.1038/nprot.2007.514 · Cited by 2,578
Hematopoiesis: An Evolving Paradigm for Stem Cell Biology
Orkin, S. H., & Zon, L. I. (2008) · Cell · doi:10.1016/j.cell.2008.01.025 · Cited by 2,570
Cited by
Best practices for single-cell analysis across modalities
Heumos, L., Schaar, A. C., Lance, C., Litinetskaya, A., Drost, F., Zappia, L., Lücken, M. D., Strobl, D. C., Henao, J., Curion, F., et al. (2023) · Nature Reviews Genetics · doi:10.1038/s41576-023-00586-w · Cited by 1,051
The technological landscape and applications of single-cell multi-omics
Baysoy, A., Bai, Z., Satija, R., & Fan, R. (2023) · Nature Reviews Molecular Cell Biology · doi:10.1038/s41580-023-00615-w · Cited by 1,037
SCENIC+: single-cell multiomic inference of enhancers and gene regulatory networks
Bravo González-Blas, C., De Winter, S., Hulselmans, G., Hecker, N., Matetovici, I., Christiaens, V., Poovathingal, S., Wouters, J., Aibar, S., & Aerts, S. (2023) · Nature Methods · doi:10.1038/s41592-023-01938-4 · Cited by 778
SCENIC+: single-cell multiomic inference of enhancers and gene regulatory networks
González-Blas, C. B., De Winter, S., Hulselmans, G., Hecker, N., Matetovici, I., Christiaens, V., Poovathingal, S., Wouters, J., Aibar, S., & Aerts, S. (2022) · doi:10.1101/2022.08.19.504505 · Cited by 778
Large-scale foundation model on single-cell transcriptomics
Hao, M., Gong, J., Zeng, X., Liu, C., Guo, Y., Cheng, X., Wang, T., Ma, J., Zhang, X., & Song, L. (2024) · Nature Methods · doi:10.1038/s41592-024-02305-7 · Cited by 628
Metadata
- Authors
- Kamimoto, K., Stringa, B., Hoffmann, C. M., Jindal, K., Solnica-Krezel, L., & Morris, S. A.
- Affiliations
- Washington University in St. Louis (US)
- Corresponding author
- Morris, Samantha A.
- ORCID
- Venue
- Nature
- Published
- 2023-02-08
- Volume
- 614
- Issue
- 7949
- Pages
- 742-751
- ISSN
- 0028-0836 (ISSN-L)
- 1476-4687
- PMC
- PMC9946838
- OpenAlex
- W4319591822
- Semantic Scholar
- 9672a8fe4712abd9e74c40d8e906c9a82f33ef39
- Cited by
- 677
- References
- 89
- Citations by year
- 2026: 226 · 2025: 198 · 2024: 146 · 2023: 80
- FWCI
- 57.74
- Type
- journal article
- Language
- en
- Open access
- hybrid
- doi.org
- Licence
- cc-by
- creativecommons.org/licenses/by/4.0 (tdm)
- creativecommons.org/licenses/by/4.0 (vor)
- Locations
- Nature (publishedVersion, cc-by)
- PubMed Central (publishedVersion, cc-by)
- PubMed (publishedVersion)
- Digital Commons@Becker (Washington University School of Medicine) (submittedVersion, cc-by)
- RePEc: Research Papers in Economics (submittedVersion)
- Funders
- National Science Foundation (1745038, 2139839)
- Silicon Valley Community Foundation (DAF2021-238797, HCA2-A-1708- 02799)
- Washington University in St. Louis
- Paul G. Allen Frontiers Group
- National Institutes of Health (5r01gm126112-05)
- Japan Society for the Promotion of Science
- National Institute of General Medical Sciences (GM126112, R01-GM126112, R35GM118179)
- Topics
- Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Gene Regulatory Network Analysis
- 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
- Keywords
- cell identity, gene regulatory networks, transcription factor perturbation, single-cell multi-omics, hematopoiesis, zebrafish embryonic development, Noto, notochord, axial mesoderm, developmental regulation, differentiation
- MeSH terms
- Animals, Cell Differentiation, Computer Simulation, Hematopoiesis, Humans, Mesoderm, Phenotype, Transcription Factors, Zebrafish, Homeodomain Proteins, Zebrafish Proteins, Embryonic Development, Mice, Gene Regulatory Networks, LIM-Homeodomain Proteins
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.
The claim verdicts this report counts are experimental: their accuracy is not yet measured.
A language model drew the claims from the paper on 2026-10-05, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-10-05 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Cited sources are checked with a language model’s help (qwen3.8-omni-flash and gpt-6-sol).
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
How to cite
Cite CoChat for the checks on this page, and the paper for its own findings.
- This page
- CoChat. (2026). Citation check: Dissecting cell identity via network inference and in silico gene perturbation. https://app.cochat.ai/papers/10.1038/s41586-022-05688-9
- The paper
- Kamimoto, K., Stringa, B., Hoffmann, C. M., Jindal, K., Solnica-Krezel, L., & Morris, S. A. (2023). Dissecting cell identity via network inference and in silico gene perturbation. Nature. https://doi.org/10.1038/s41586-022-05688-9
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.