Academic Journal
A systematic comparison of single-cell perturbation response prediction models.
| Τίτλος: | A systematic comparison of single-cell perturbation response prediction models. |
|---|---|
| Συγγραφείς: | Li L; GMU-GIBH Joint School of Life Sciences, Guangzhou Medical University, Guangzhou, China.; Guangzhou National Laboratory, Guangzhou, China., You Y; Guangzhou National Laboratory, Guangzhou, China., Fu Y; Guangzhou National Laboratory, Guangzhou, China., Liao W; The University of Hong Kong, Hong Kong, China., Fan X; School of Life Sciences, Westlake University, Hangzhou, China.; Research Center for Industries of the Future, Westlake University, Hangzhou, China.; Westlake Laboratory of Life Sciences and Biomedicine, Hangzhou, China.; Westlake Institute for Advanced Study, Hangzhou, China., Lu S; Guangzhou National Laboratory, Guangzhou, China., Cao Y; Guangzhou National Laboratory, Guangzhou, China., Li B; Guangzhou National Laboratory, Guangzhou, China., Ren W; Guangzhou National Laboratory, Guangzhou, China., Kong J; Renaissance Era Innovation Technology Co., Ltd., Beijing, China., Zheng S; Global Institute of Future Technology, Shanghai Jiao Tong University, Shanghai, China., Chen J; Guangzhou National Laboratory, Guangzhou, China.; State Key Laboratory of Respiratory Disease, National Clinical Research Center for Respiratory Disease, Guangzhou Institute of Respiratory Health, The First Affiliated Hospital of Guangzhou Medical University, Guangzhou, China., Liu X; School of Life Sciences, Westlake University, Hangzhou, China.; Research Center for Industries of the Future, Westlake University, Hangzhou, China.; Westlake Laboratory of Life Sciences and Biomedicine, Hangzhou, China.; Westlake Institute for Advanced Study, Hangzhou, China., Tian L; GMU-GIBH Joint School of Life Sciences, Guangzhou Medical University, Guangzhou, China.; Guangzhou National Laboratory, Guangzhou, China. |
| Πηγή: | Science advances [Sci Adv] 2026 Sep 11; Vol. 12 (37), pp. eaed3414. Date of Electronic Publication: 2026 Sep 09. |
| Τύπος έκδοσης: | Journal Article; Comparative Study |
| Γλώσσα: | English |
| Στοιχεία περιοδικού: | Publisher: American Association for the Advancement of Science Country of Publication: United States NLM ID: 101653440 Publication Model: Print-Electronic Cited Medium: Internet ISSN: 2375-2548 (Electronic) Linking ISSN: 23752548 NLM ISO Abbreviation: Sci Adv Subsets: MEDLINE |
| Imprint Name(s): | Original Publication: Washington, DC : American Association for the Advancement of Science, [2015]- |
| Ιατρικοί όροι (MeSH): | Single-Cell Analysis*/methods , Computational Biology*/methods , Gene Expression Regulation*, Humans ; Animals ; Single-Cell Gene Expression Analysis ; Gene Expression Profiling |
| Περίληψη: | Predicting single-cell transcriptional responses to perturbations is central to dissecting gene regulation and accelerating therapeutic design, yet the field lacks a rigorous, task-spanning assessment of model behavior. We present a large-scale benchmark of 13 representative methods and baselines across 25 datasets spanning diverse perturbation modalities and species, including two primary immune-cell drug-response resources. We evaluated three core tasks-generalization to unseen single-gene perturbations, prediction of combinatorial interactions, and transfer across cell types-using 24 metrics covering expression-level accuracy, relative changes, differential expression (DE) recovery, and distributional similarity. Across tasks, performance depended strongly on perturbation effect size and evaluation perspective: Expression-level agreement was the highest for small-effect perturbations resembling controls, whereas delta- and DE-based metrics improved with larger effects, providing clearer signals. Models shared a conservative bias, with fine-tuned foundation models compressing variance and underestimating synergistic effects in combinations. PerturbNet showed superior recovery of DE signatures in Tasks 1 and 2, while no method consistently generalized across cell types in Task 3, where biological consistency dominated outcomes. This benchmark establishes current methodological limits, clarifies that different metrics probe distinct biological signals rather than redundant summaries of the same prediction problem, and provides a foundation for developing virtual-cell models that more faithfully capture heterogeneous perturbation responses. |
| References: | Cell. 2025 Jun 26;188(13):3370-3374. (PMID: 40578317) Nat Commun. 2026 Apr 24;17(1):. (PMID: 42031729) Mol Syst Biol. 2023 Jun 12;19(6):e11517. (PMID: 37154091) Cell Syst. 2021 Jun 16;12(6):522-537. (PMID: 34139164) Nat Neurosci. 2021 Jul;24(7):1020-1034. (PMID: 34031600) Nat Rev Microbiol. 2020 Dec;18(12):731-743. (PMID: 32958892) Nat Genet. 2025 May;57(5):1228-1237. (PMID: 40259084) Nat Methods. 2024 Aug;21(8):1481-1491. (PMID: 38844628) Nature. 2018 Nov;563(7730):197-202. (PMID: 30356220) Nat Methods. 2026 Feb;23(2):451-464. (PMID: 41381899) Nat Methods. 2025 Aug;22(8):1657-1661. (PMID: 40759747) Bioinformatics. 2020 Dec 30;36(Suppl_2):i610-i617. (PMID: 33381839) Nat Biotechnol. 2024 Nov;42(11):1678-1683. (PMID: 38225466) Adv Sci (Weinh). 2024 Aug;11(32):e2404845. (PMID: 39031820) Nat Methods. 2023 Nov;20(11):1759-1768. (PMID: 37770709) Nucleic Acids Res. 2021 Dec 2;49(21):e122. (PMID: 34500471) Cell. 2022 Jul 7;185(14):2559-2575.e28. (PMID: 35688146) Cell. 2024 Aug 22;187(17):4520-4545. (PMID: 39178831) Nat Methods. 2017 Mar;14(3):297-301. (PMID: 28099430) Patterns (N Y). 2023 Aug 11;4(8):100817. (PMID: 37602218) Science. 2019 Aug 23;365(6455):786-793. (PMID: 31395745) Cell. 2016 Dec 15;167(7):1853-1866.e17. (PMID: 27984732) Nat Genet. 2021 Mar;53(3):322-331. (PMID: 33649593) Science. 2020 Jan 3;367(6473):45-51. (PMID: 31806696) Brief Funct Genomics. 2018 Jul 1;17(4):246-254. (PMID: 29342231) Bioinformatics. 2022 Jun 27;38(13):3377-3384. (PMID: 35639705) Bioinformatics. 2025 Jun 2;41(6):. (PMID: 40407144) Nat Biotechnol. 2024 Jun;42(6):927-935. (PMID: 37592036) Mol Syst Biol. 2025 Aug;21(8):960-982. (PMID: 40640612) Nat Methods. 2024 Mar;21(3):531-540. (PMID: 38279009) Mol Syst Biol. 2020 Sep;16(9):e9198. (PMID: 32975352) Nat Commun. 2023 Oct 6;14(1):6245. (PMID: 37803001) Nat Biotechnol. 2026 Jun;44(6):1050-1059. (PMID: 40854979) Science. 2020 Feb 14;367(6479):. (PMID: 31974159) Cell Syst. 2025 Jan 15;16(1):101161. (PMID: 39689711) Cell. 2024 Nov 14;187(23):6421-6423. (PMID: 39547206) Bioinformatics. 2024 May 2;40(5):. (PMID: 38625746) Cell. 2024 Dec 12;187(25):7045-7063. (PMID: 39672099) Nat Genet. 2021 Mar;53(3):332-341. (PMID: 33649592) Nat Commun. 2019 May 20;10(1):2233. (PMID: 31110232) Nat Methods. 2019 Aug;16(8):715-721. (PMID: 31363220) Cell. 2025 Aug 21;188(17):4790-4809.e22. (PMID: 40513557) Genome Biol. 2025 Apr 18;26(1):101. (PMID: 40251685) Cell. 2016 Dec 15;167(7):1867-1882.e21. (PMID: 27984733) Nat Methods. 2021 Jun;18(6):635-642. (PMID: 34059827) Nat Methods. 2024 Aug;21(8):1470-1480. (PMID: 38409223) Nat Biotechnol. 2018 Jan;36(1):89-94. (PMID: 29227470) Nat Genet. 2015 Aug;47(8):856-60. (PMID: 26121088) Basic Res Cardiol. 2023 Feb 7;118(1):7. (PMID: 36750503) Nature. 2023 Jun;618(7965):616-624. (PMID: 37258680) Nature. 2025 Jan;637(8048):1065-1071. (PMID: 39566552) Comput Struct Biotechnol J. 2024 Apr 25;23:1886-1896. (PMID: 38721585) Nature. 2017 Nov 16;551(7680):333-339. (PMID: 29144463) Genome Biol. 2025 Nov 17;26(1):388. (PMID: 41250104) Nat Methods. 2026 Jan;23(1):65-77. (PMID: 41184550) |
| Entry Date(s): | Date Created: 20260909 Date Completed: 20260909 Latest Revision: 20260914 |
| Update Code: | 20260914 |
| PubMed Central ID: | PMC13557097 |
| DOI: | 10.1126/sciadv.aed3414 |
| PMID: | 42715312 |
| Βάση Δεδομένων: | MEDLINE |
καταχωρήστε σχόλιο πρώτοι!