{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"institution":[{"name":"bioRxiv"}],"indexed":{"date-parts":[[2026,9,30]],"date-time":"2026-09-30T03:56:20Z","timestamp":1790740580941,"version":"4.1.0"},"posted":{"date-parts":[[2026,9,26]]},"group-title":"Bioinformatics","reference-count":32,"publisher":"openRxiv","license":[{"start":{"date-parts":[[2026,9,26]],"date-time":"2026-09-26T00:00:00Z","timestamp":1790380800000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"id":[{"id":"https:\/\/ror.org\/001d55x84","id-type":"ROR","asserted-by":"publisher"}]},{"id":[{"id":"https:\/\/ror.org\/00yjd3n13","id-type":"ROR","asserted-by":"publisher"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"accepted":{"date-parts":[[2026,9,26]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                <jats:p>Genetic perturbations can reshape cell populations by altering the relative abundance of specific cell types and states within the profiled population, including increases, decreases and states that become detectable after perturbation. Pooled single-cell screens, such as Perturb-seq, measure such responses at scale. How-ever, only a small fraction of possible perturbations can be tested experimentally. A central challenge is therefore to predict compositional shifts induced by unseen perturbations. Many perturbation-prediction methods do not directly optimize for this outcome; instead, they predict gene-expression responses and infer cell-type and cell-state composition downstream. Surprisingly, we find that even models that accurately predict perturbation-induced changes in average gene expression perform poorly at forecasting these compositional shifts. To address this gap, we present Pop-Corn, a method that directly predicts how a perturbation reshapes cell-type composition without reconstructing gene expression. In the primary T-cell benchmark, Pop-Corn predicted the overall cell-state composition of held-out perturbations more accurately than the evaluated expression-prediction pipelines, while better preserving the diversity of observed cell states. We further extend Pop-Corn to intact tissue, where it predicts perturbation-induced cell-type proportion changes in local cellular neighborhoods and uses attention patterns to generate hypotheses about context-dependent cellular interactions. Retrospective virtual screens support the use of Pop-Corn to prioritize perturbations for experimental follow-up according to their predicted effects on cell-state composition.<\/jats:p>","DOI":"10.64898\/2026.09.25.754560","type":"posted-content","created":{"date-parts":[[2026,9,26]],"date-time":"2026-09-26T16:50:19Z","timestamp":1790441419000},"source":"Crossref","is-referenced-by-count":0,"title":["Pop-Corn: Predicting Perturbation Phenotype Effects Across Single-Cell and Spatial Contexts"],"prefix":"10.64898","author":[{"given":"Jiapeng","family":"Chen","sequence":"first","affiliation":[{"name":"Computational Biology and Bioinformatics Program, Yale University, New Haven, CT, USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yan","family":"Cui","sequence":"additional","affiliation":[{"name":"Whitehead Institute for Biomedical Research, Cambridge, MA, USA"},{"name":"Department of Bioengineering, University of Pennsylvania, Philadelphia, PA, USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yanjun","family":"Shao","sequence":"additional","affiliation":[{"name":"Computational Biology and Bioinformatics Program, Yale University, New Haven, CT, USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Na","family":"Sun","sequence":"additional","affiliation":[{"name":"Whitehead Institute for Biomedical Research, Cambridge, MA, USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Mar\u00eda Rodr\u00edguez","family":"Mart\u00ednez","sequence":"additional","affiliation":[{"name":"Biomedical Informatics and Data Science, Yale School of Medicine, New Haven, CT, USA"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"54368","reference":[{"key":"2026092920051188000_2026.09.25.754560v1.1","doi-asserted-by":"publisher","DOI":"10.1016\/j.cell.2016.11.038"},{"key":"2026092920051188000_2026.09.25.754560v1.2","doi-asserted-by":"publisher","DOI":"10.1016\/j.cell.2022.05.013"},{"key":"2026092920051188000_2026.09.25.754560v1.3","doi-asserted-by":"publisher","DOI":"10.1038\/s41587-023-01905-6"},{"key":"2026092920051188000_2026.09.25.754560v1.4","doi-asserted-by":"publisher","DOI":"10.1038\/s41592-024-02201-0"},{"key":"2026092920051188000_2026.09.25.754560v1.5","doi-asserted-by":"crossref","unstructured":"Klein, D. et al. CellFlow enables generative single-cell phenotype modeling with flow matching. bioRxiv 2025.04.11.648220 (2025).","DOI":"10.1101\/2025.04.11.648220"},{"key":"2026092920051188000_2026.09.25.754560v1.6","doi-asserted-by":"crossref","unstructured":"Ji, Y. et al. Scalable and universal prediction of cellular phenotypes enables in silico experiments. bioRxiv 2024.08.12.607533 (2024).","DOI":"10.1101\/2024.08.12.607533"},{"key":"2026092920051188000_2026.09.25.754560v1.7","doi-asserted-by":"crossref","unstructured":"He, C. , Zhang, J. , Dahleh, M. & Uhler, C. MORPH predicts the single-cell outcome of genetic perturbations across conditions and data modalities. bioRxiv 2025.06.27.661992 (2025).","DOI":"10.1101\/2025.06.27.661992"},{"key":"2026092920051188000_2026.09.25.754560v1.8","doi-asserted-by":"publisher","DOI":"10.1038\/s41592-023-01969-x"},{"key":"2026092920051188000_2026.09.25.754560v1.9","doi-asserted-by":"publisher","DOI":"10.1016\/j.cell.2021.12.045"},{"key":"2026092920051188000_2026.09.25.754560v1.10","doi-asserted-by":"publisher","DOI":"10.1126\/science.abj4008"},{"key":"2026092920051188000_2026.09.25.754560v1.11","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","year":"2023","journal-title":"Nat. Mach. Intell"},{"key":"2026092920051188000_2026.09.25.754560v1.12","doi-asserted-by":"publisher","DOI":"10.15252\/msb.202211517"},{"key":"2026092920051188000_2026.09.25.754560v1.13","doi-asserted-by":"publisher","DOI":"10.1038\/s44320-025-00131-3"},{"key":"2026092920051188000_2026.09.25.754560v1.14","doi-asserted-by":"crossref","unstructured":"Vi\u00f1as Torn\u00e9, R. et al. Systema: a framework for evaluating genetic perturbation response prediction beyond systematic variation. Nat. Biotechnol. 1\u201310 (2025).","DOI":"10.1038\/s41587-025-02777-8"},{"key":"2026092920051188000_2026.09.25.754560v1.15","doi-asserted-by":"crossref","first-page":"1657","DOI":"10.1038\/s41592-025-02772-6","article-title":"Deep-learning-based gene perturbation effect prediction does not yet outperform simple linear baselines","volume":"22","year":"2025","journal-title":"Nat. Methods"},{"key":"2026092920051188000_2026.09.25.754560v1.16","unstructured":"Mejia, G. M. et al. Diversity by design: Addressing mode collapse improves scRNA-seq perturbation modeling on well-calibrated metrics. Poster presented at the 2nd Generative AI for Biology Workshop, 42nd International Conference on Machine Learning (2025)."},{"key":"2026092920051188000_2026.09.25.754560v1.17","doi-asserted-by":"publisher","DOI":"10.1126\/science.aax4438"},{"key":"2026092920051188000_2026.09.25.754560v1.18","doi-asserted-by":"crossref","unstructured":"Zhu, R. et al. Genome-scale Perturb-seq in primary human CD4+ T cells maps context-specific regulators of T cell programs and human immune traits. bioRxiv 2025.12.23.696273 (2025).","DOI":"10.64898\/2025.12.23.696273"},{"key":"2026092920051188000_2026.09.25.754560v1.19","doi-asserted-by":"crossref","first-page":"4790","DOI":"10.1016\/j.cell.2025.05.022","article-title":"Perturb-multimodal: A platform for pooled genetic screens with imaging and sequencing in intact mammalian tissue","volume":"188","year":"2025","journal-title":"Cell"},{"key":"2026092920051188000_2026.09.25.754560v1.20","doi-asserted-by":"crossref","unstructured":"Hong, J. et al. AMICI: Attention mechanism interpretation of cell\u2013cell interactions. bioRxiv 2025.09.22.677860 (2025).","DOI":"10.1101\/2025.09.22.677860"},{"key":"2026092920051188000_2026.09.25.754560v1.21","doi-asserted-by":"publisher","DOI":"10.1038\/s41587-021-01001-7"},{"key":"2026092920051188000_2026.09.25.754560v1.22","doi-asserted-by":"publisher","DOI":"10.1126\/science.abl5197"},{"key":"2026092920051188000_2026.09.25.754560v1.23","doi-asserted-by":"publisher","DOI":"10.1038\/s41467-023-43105-5"},{"key":"2026092920051188000_2026.09.25.754560v1.24","doi-asserted-by":"crossref","unstructured":"Heo, B. , Park, S. , Han, D. & Yun, S. Rotary position embedding for vision transformer. European Conference on Computer Vision (ECCV) (2024).","DOI":"10.1007\/978-3-031-72684-2_17"},{"key":"2026092920051188000_2026.09.25.754560v1.25","doi-asserted-by":"publisher","DOI":"10.1242\/dev.104539"},{"key":"2026092920051188000_2026.09.25.754560v1.26","unstructured":"Ilse, M. , Tomczak, J. & Welling, M. Dy, J. & Krause, A. (eds) Attention-based deep multiple instance learning. (eds Dy, J. & Krause, A. ) Proceedings of the 35th International Conference on Machine Learning, Vol. 80 of Proceedings of Machine Learning Research, 2127\u20132136 (PMLR, 2018)."},{"key":"2026092920051188000_2026.09.25.754560v1.27","doi-asserted-by":"publisher","DOI":"10.1038\/s41592-023-02040-5"},{"key":"2026092920051188000_2026.09.25.754560v1.28","doi-asserted-by":"crossref","first-page":"555","DOI":"10.1038\/s41551-020-00682-w","article-title":"Data-efficient and weakly supervised computational pathology on whole-slide images","volume":"5","year":"2021","journal-title":"Nat. Biomed. Eng"},{"key":"2026092920051188000_2026.09.25.754560v1.29","doi-asserted-by":"crossref","unstructured":"Litinetskaya, A. , Hediyeh-zadeh, S. , Moinfar, A. A. , Lotfollahi, M. & Theis, F. J. Weakly supervised learning uncovers phenotypic signatures in single-cell data. bioRxiv 2024.07.29.605625 (2024).","DOI":"10.1101\/2024.07.29.605625"},{"key":"2026092920051188000_2026.09.25.754560v1.30","doi-asserted-by":"publisher","DOI":"10.1126\/science.ade2574"},{"key":"2026092920051188000_2026.09.25.754560v1.31","doi-asserted-by":"publisher","DOI":"10.5281\/zenodo.5784651"},{"key":"2026092920051188000_2026.09.25.754560v1.32","doi-asserted-by":"publisher","DOI":"10.1186\/s13059-017-1382-0"}],"container-title":[],"original-title":[],"link":[{"URL":"https:\/\/syndication.highwire.org\/content\/doi\/10.64898\/2026.09.25.754560","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,9,30]],"date-time":"2026-09-30T03:05:24Z","timestamp":1790737524000},"score":1,"resource":{"primary":{"URL":"http:\/\/biorxiv.org\/lookup\/doi\/10.64898\/2026.09.25.754560"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,9,26]]},"references-count":32,"URL":"https:\/\/doi.org\/10.64898\/2026.09.25.754560","relation":{},"subject":[],"published":{"date-parts":[[2026,9,26]]},"subtype":"preprint"}}