{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,2,10]],"date-time":"2026-02-10T20:46:17Z","timestamp":1770756377942,"version":"3.50.0"},"update-to":[{"DOI":"10.1371\/journal.pcbi.1013194","type":"new_version","label":"New version","source":"publisher","updated":{"date-parts":[[2026,2,9]],"date-time":"2026-02-09T00:00:00Z","timestamp":1770595200000}}],"reference-count":40,"publisher":"Public Library of Science (PLoS)","issue":"2","license":[{"start":{"date-parts":[[2026,2,2]],"date-time":"2026-02-02T00:00:00Z","timestamp":1769990400000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"DOI":"10.13039\/100013114","name":"Broad Institute","doi-asserted-by":"publisher","award":["Eric and Wendy Schmidt Center"],"award-info":[{"award-number":["Eric and Wendy Schmidt Center"]}],"id":[{"id":"10.13039\/100013114","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100017567","name":"Apple","doi-asserted-by":"publisher","award":["AI\/ML PhD fellowship"],"award-info":[{"award-number":["AI\/ML PhD fellowship"]}],"id":[{"id":"10.13039\/100017567","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100008460","name":"National Center for Complementary and Integrative Health","doi-asserted-by":"publisher","award":["1DP2AT012345"],"award-info":[{"award-number":["1DP2AT012345"]}],"id":[{"id":"10.13039\/100008460","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100000006","name":"Office of Naval Research","doi-asserted-by":"publisher","award":["N00014-22-1-2116"],"award-info":[{"award-number":["N00014-22-1-2116"]}],"id":[{"id":"10.13039\/100000006","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100000006","name":"Office of Naval Research","doi-asserted-by":"publisher","award":["N00014-24-1-2687"],"award-info":[{"award-number":["N00014-24-1-2687"]}],"id":[{"id":"10.13039\/100000006","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100000015","name":"U.S. Department of Energy","doi-asserted-by":"publisher","award":["DE-SC0023187"],"award-info":[{"award-number":["DE-SC0023187"]}],"id":[{"id":"10.13039\/100000015","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100013114","name":"Broad Institute","doi-asserted-by":"publisher","id":[{"id":"10.13039\/100013114","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100000893","name":"Simons Foundation","doi-asserted-by":"publisher","award":["Simons Investigator Award"],"award-info":[{"award-number":["Simons Investigator Award"]}],"id":[{"id":"10.13039\/100000893","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["www.ploscompbiol.org"],"crossmark-restriction":false},"short-container-title":["PLoS Comput Biol"],"abstract":"<jats:p>Advances in sequencing technologies have enhanced the understanding of gene regulation in cells. In particular, Perturb-seq has enabled high-resolution profiling of the transcriptomic response to genetic perturbations at the single-cell level. This understanding has implications in functional genomics and potentially for identifying therapeutic targets. Various computational models have been developed to predict perturbational effects. While deep learning models excel at interpolating observed perturbational data, they tend to overfit in the lack of enough data and may not generalize well to unseen perturbations. In contrast, mechanistic models, such as linear causal models based on gene regulatory networks, hold greater potential for extrapolation, as they encapsulate regulatory information that can predict responses to unseen perturbations. However, their application has been limited to small studies due to overly simplistic assumptions, making them less effective in handling noisy, large-scale single-cell data. We propose a hybrid approach that combines a mechanistic causal model with variational deep learning, termed Single Cell Causal Variational Autoencoder (SCCVAE). The mechanistic model employs a learned regulatory network to represent perturbational changes as shift interventions that propagate through the learned network. SCCVAE integrates this mechanistic causal model into a variational autoencoder, generating rich, comprehensive transcriptomic responses. Our results indicate that SCCVAE exhibits superior performance over current state-of-the-art baselines for extrapolating to predict unseen perturbational responses. Additionally, for the observed perturbations, the latent space learned by SCCVAE allows for the identification of functional perturbation modules and simulation of single-gene knockdown experiments of varying penetrance, presenting a robust tool for interpreting and interpolating perturbational responses at the single-cell level.<\/jats:p>","DOI":"10.1371\/journal.pcbi.1013194","type":"journal-article","created":{"date-parts":[[2026,2,2]],"date-time":"2026-02-02T18:48:13Z","timestamp":1770058093000},"page":"e1013194","update-policy":"https:\/\/doi.org\/10.1371\/journal.pcbi.corrections_policy","source":"Crossref","is-referenced-by-count":0,"title":["Learning genetic perturbation effects with variational causal inference"],"prefix":"10.1371","volume":"22","author":[{"given":"Emily","family":"Liu","sequence":"first","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-9039-6843","authenticated-orcid":true,"given":"Jiaqi","family":"Zhang","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-7008-0216","authenticated-orcid":true,"given":"Caroline","family":"Uhler","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"340","published-online":{"date-parts":[[2026,2,2]]},"reference":[{"issue":"1","key":"pcbi.1013194.ref001","doi-asserted-by":"crossref","first-page":"9","DOI":"10.1038\/s43586-021-00093-4","article-title":"High-content CRISPR screening","volume":"2","author":"C Bock","year":"2022","journal-title":"Nat Rev Methods Primers."},{"issue":"6399","key":"pcbi.1013194.ref002","doi-asserted-by":"crossref","first-page":"285","DOI":"10.1126\/science.aao0932","article-title":"Domain-focused CRISPR screen identifies HRI as a fetal hemoglobin regulator in human erythroid cells","volume":"361","author":"JD Grevet","year":"2018","journal-title":"Science."},{"issue":"5","key":"pcbi.1013194.ref003","doi-asserted-by":"crossref","DOI":"10.1016\/j.neuron.2019.09.003","article-title":"CRISPR-Cas9 Screens Identify the RNA Helicase DDX3X as a Repressor of C9ORF72 (GGGGCC)n Repeat-Associated Non-AUG Translation","volume":"104","author":"W Cheng","year":"2019","journal-title":"Neuron."},{"issue":"6538","key":"pcbi.1013194.ref004","doi-asserted-by":"crossref","DOI":"10.1126\/science.abd0875","article-title":"QSER1 protects DNA methylation valleys from de novo methylation","volume":"372","author":"G Dixon","year":"2021","journal-title":"Science."},{"issue":"6166","key":"pcbi.1013194.ref005","doi-asserted-by":"crossref","first-page":"80","DOI":"10.1126\/science.1246981","article-title":"Genetic screens in human cells using the CRISPR-Cas9 system","volume":"343","author":"T Wang","year":"2014","journal-title":"Science."},{"issue":"6","key":"pcbi.1013194.ref006","doi-asserted-by":"crossref","first-page":"1515","DOI":"10.1016\/j.cell.2015.11.015","article-title":"High-resolution CRISPR screens reveal fitness genes and genotype-specific cancer liabilities","volume":"163","author":"T Hart","year":"2015","journal-title":"Cell."},{"issue":"7","key":"pcbi.1013194.ref007","doi-asserted-by":"crossref","DOI":"10.1016\/j.cell.2016.11.038","article-title":"Perturb-Seq: dissecting molecular circuits with scalable single-cell RNA profiling of pooled genetic screens","volume":"167","author":"A Dixit","year":"2016","journal-title":"Cell."},{"issue":"7","key":"pcbi.1013194.ref008","doi-asserted-by":"crossref","DOI":"10.1016\/j.cell.2016.11.048","article-title":"A multiplexed single-cell CRISPR screening platform enables systematic dissection of the unfolded protein response","volume":"167","author":"B Adamson","year":"2016","journal-title":"Cell."},{"issue":"14","key":"pcbi.1013194.ref009","doi-asserted-by":"crossref","DOI":"10.1016\/j.cell.2022.05.013","article-title":"Mapping information-rich genotype-phenotype landscapes with genome-scale Perturb-seq","volume":"185","author":"JM Replogle","year":"2022","journal-title":"Cell."},{"issue":"6","key":"pcbi.1013194.ref010","doi-asserted-by":"crossref","DOI":"10.15252\/msb.202211517","article-title":"Predicting cellular responses to complex perturbations in high-throughput screens","volume":"19","author":"M Lotfollahi","year":"2023","journal-title":"Mol Syst Biol."},{"key":"pcbi.1013194.ref011","article-title":"Perturbnet predicts single-cell responses to unseen chemical and genetic perturbations","author":"H Yu","year":"2022","journal-title":"BioRxiv."},{"issue":"6","key":"pcbi.1013194.ref012","doi-asserted-by":"crossref","first-page":"927","DOI":"10.1038\/s41587-023-01905-6","article-title":"Predicting transcriptional outcomes of novel multigene perturbations with GEARS","volume":"42","author":"Y Roohani","year":"2024","journal-title":"Nat Biotechnol."},{"key":"pcbi.1013194.ref013","unstructured":"Lopez R, Tagasovska N, Ra S, Cho K, Pritchard J, Regev A. Learning causal representations of single cells via sparse mechanism shift modeling. In: Conference on Causal Learning and Reasoning. PMLR; 2023. p. 662\u201391."},{"key":"pcbi.1013194.ref014","unstructured":"Wu Y, Barton RA, Wang Z, Ioannidis VN, De Donno C, Price LC. Predicting cellular responses with variational causal inference and refined relational information. arXiv preprint 2022. https:\/\/arxiv.org\/abs\/221000116"},{"issue":"1","key":"pcbi.1013194.ref015","doi-asserted-by":"crossref","first-page":"6775","DOI":"10.1038\/s41467-025-62008-1","article-title":"Causal disentanglement for single-cell representations and controllable counterfactual generation","volume":"16","author":"Y Gao","year":"2025","journal-title":"Nat Commun."},{"key":"pcbi.1013194.ref016","doi-asserted-by":"crossref","unstructured":"Aliee H, Kapl F, Hediyeh-Zadeh S, Theis FJ. Conditionally invariant representation learning for disentangling cellular heterogeneity. arXiv preprint 2023. https:\/\/arxiv.org\/abs\/230700558","DOI":"10.1101\/2024.12.06.627196"},{"key":"pcbi.1013194.ref017","article-title":"A supervised contrastive framework for learning disentangled representations of cell perturbation data","author":"X Tu","year":"2024","journal-title":"bioRxiv."},{"issue":"17","key":"pcbi.1013194.ref018","doi-asserted-by":"crossref","first-page":"4520","DOI":"10.1016\/j.cell.2024.07.035","article-title":"Toward a foundation model of causal cell and tissue biology with a Perturbation Cell and Tissue Atlas","volume":"187","author":"JE Rood","year":"2024","journal-title":"Cell."},{"issue":"7965","key":"pcbi.1013194.ref019","doi-asserted-by":"crossref","first-page":"616","DOI":"10.1038\/s41586-023-06139-9","article-title":"Transfer learning enables predictions in network biology","volume":"618","author":"CV Theodoris","year":"2023","journal-title":"Nature."},{"issue":"8","key":"pcbi.1013194.ref020","doi-asserted-by":"crossref","first-page":"1470","DOI":"10.1038\/s41592-024-02201-0","article-title":"scGPT: toward building a foundation model for single-cell multi-omics using generative AI","volume":"21","author":"H Cui","year":"2024","journal-title":"Nat Methods."},{"key":"pcbi.1013194.ref021","article-title":"Deep learning-based predictions of gene perturbation effects do not yet outperform simple linear methods","author":"C Ahlmann-Eltze","year":"2024","journal-title":"BioRxiv."},{"key":"pcbi.1013194.ref022","first-page":"6859","article-title":"Supervised training of conditional monge maps","volume":"35","author":"C Bunne","year":"2022","journal-title":"Advances in Neural Information Processing Systems."},{"key":"pcbi.1013194.ref023","article-title":"Identifiability guarantees for causal disentanglement from soft interventions","volume":"36","author":"J Zhang","year":"2024","journal-title":"Advances in Neural Information Processing Systems."},{"key":"pcbi.1013194.ref024","unstructured":"Gaudelet T, Del Vecchio A, Carrami EM, Cudini J, Kapourani CA, Uhler C, et al. Season combinatorial intervention predictions with Salt & Peper. arXiv preprint 2024. arXiv:240416907"},{"issue":"7949","key":"pcbi.1013194.ref025","doi-asserted-by":"crossref","first-page":"742","DOI":"10.1038\/s41586-022-05688-9","article-title":"Dissecting cell identity via network inference and in silico gene perturbation","volume":"614","author":"K Kamimoto","year":"2023","journal-title":"Nature."},{"issue":"10","key":"pcbi.1013194.ref026","doi-asserted-by":"crossref","first-page":"1066","DOI":"10.1038\/s42256-023-00719-0","article-title":"Active learning for optimal intervention design in causal models","volume":"5","author":"J Zhang","year":"2023","journal-title":"Nat Mach Intell."},{"issue":"3","key":"pcbi.1013194.ref027","article-title":"SERGIO: a single-cell expression simulator guided by gene regulatory networks","volume":"11","author":"P Dibaeinia","year":"2020","journal-title":"Cell Syst."},{"key":"pcbi.1013194.ref028","doi-asserted-by":"crossref","unstructured":"Spirtes P, Glymour C, Scheines R. Causation, prediction, and search. MIT Press; 2001.","DOI":"10.7551\/mitpress\/1754.001.0001"},{"key":"pcbi.1013194.ref029","doi-asserted-by":"crossref","unstructured":"Pearl J. Causality. Cambridge University Press; 2009.","DOI":"10.1017\/CBO9780511803161"},{"issue":"12","key":"pcbi.1013194.ref030","doi-asserted-by":"crossref","first-page":"1053","DOI":"10.1038\/s41592-018-0229-2","article-title":"Deep generative modeling for single-cell transcriptomics","volume":"15","author":"R Lopez","year":"2018","journal-title":"Nat Methods."},{"issue":"8","key":"pcbi.1013194.ref031","doi-asserted-by":"crossref","first-page":"715","DOI":"10.1038\/s41592-019-0494-8","article-title":"scGen predicts single-cell perturbation responses","volume":"16","author":"M Lotfollahi","year":"2019","journal-title":"Nat Methods."},{"key":"pcbi.1013194.ref032","unstructured":"Richardson TS. A discovery algorithm for directed cyclic graphs. arXiv preprint 2013. https:\/\/arxiv.org\/abs\/1302.3599"},{"key":"pcbi.1013194.ref033","unstructured":"Mooij JM, Claassen T. Constraint-based causal discovery using partial ancestral graphs in the presence of cycles. In: Conference on Uncertainty in Artificial Intelligence. PMLR; 2020. p. 1159\u201368."},{"issue":"1","key":"pcbi.1013194.ref034","doi-asserted-by":"crossref","first-page":"33","DOI":"10.1007\/s41060-018-0158-2","article-title":"A constraint-based algorithm for causal discovery with cycles, latent variables and selection bias","volume":"8","author":"EV Strobl","year":"2018","journal-title":"Int J Data Sci Anal."},{"key":"pcbi.1013194.ref035","unstructured":"Lacerda G, Spirtes PL, Ramsey J, Hoyer PO. Discovering cyclic causal models by independent components analysis. arXiv preprint 2012. https:\/\/arxiv.org\/abs\/1206.3273"},{"key":"pcbi.1013194.ref036","article-title":"On causal discovery with cyclic additive noise models","volume":"24","author":"JM Mooij","year":"2011","journal-title":"Advances in Neural Information Processing Systems."},{"key":"pcbi.1013194.ref037","unstructured":"Kingma DP. Auto-encoding variational bayes. arXiv preprint 2013. https:\/\/arxiv.org\/abs\/1312.6114"},{"issue":"1","key":"pcbi.1013194.ref038","first-page":"723","article-title":"A kernel two-sample test","volume":"13","author":"A Gretton","year":"2012","journal-title":"The Journal of Machine Learning Research."},{"issue":"14","key":"pcbi.1013194.ref039","doi-asserted-by":"crossref","DOI":"10.1016\/j.cell.2022.05.013","article-title":"Mapping information-rich genotype-phenotype landscapes with genome-scale Perturb-seq","volume":"185","author":"JM Replogle","year":"2022","journal-title":"Cell."},{"issue":"4","key":"pcbi.1013194.ref040","doi-asserted-by":"crossref","first-page":"795","DOI":"10.1093\/biomet\/asaa104","article-title":"Consistency guarantees for greedy permutation-based causal inference algorithms","volume":"108","author":"L Solus","year":"2021","journal-title":"Biometrika."}],"updated-by":[{"DOI":"10.1371\/journal.pcbi.1013194","type":"new_version","label":"New version","source":"publisher","updated":{"date-parts":[[2026,2,9]],"date-time":"2026-02-09T00:00:00Z","timestamp":1770595200000}}],"container-title":["PLOS Computational Biology"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/dx.plos.org\/10.1371\/journal.pcbi.1013194","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,2,9]],"date-time":"2026-02-09T18:59:49Z","timestamp":1770663589000},"score":1,"resource":{"primary":{"URL":"https:\/\/dx.plos.org\/10.1371\/journal.pcbi.1013194"}},"subtitle":[],"editor":[{"given":"Philipp Martin","family":"Altrock","sequence":"first","affiliation":[],"role":[{"role":"editor","vocabulary":"crossref"}]}],"short-title":[],"issued":{"date-parts":[[2026,2,2]]},"references-count":40,"journal-issue":{"issue":"2","published-online":{"date-parts":[[2026,2,2]]}},"URL":"https:\/\/doi.org\/10.1371\/journal.pcbi.1013194","relation":{},"ISSN":["1553-7358"],"issn-type":[{"value":"1553-7358","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,2,2]]}}}