{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"institution":[{"name":"bioRxiv"}],"indexed":{"date-parts":[[2026,8,17]],"date-time":"2026-08-17T19:35:26Z","timestamp":1786995326624,"version":"build-2736575974"},"posted":{"date-parts":[[2026,8,13]]},"group-title":"Bioinformatics","reference-count":48,"publisher":"openRxiv","license":[{"start":{"date-parts":[[2026,8,13]],"date-time":"2026-08-13T00:00:00Z","timestamp":1786579200000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/creativecommons.org\/licenses\/by-nc-nd\/4.0\/"}],"funder":[{"award":["U54CA209975"],"award-info":[{"award-number":["U54CA209975"]}],"id":[{"id":"https:\/\/ror.org\/040gcmg81","id-type":"ROR","asserted-by":"publisher"}]},{"DOI":"10.13039\/100000057","name":"National Institute of General Medical Sciences","doi-asserted-by":"crossref","award":["5T32GM132083"],"award-info":[{"award-number":["5T32GM132083"]}],"id":[{"id":"10.13039\/100000057","id-type":"DOI","asserted-by":"crossref"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"accepted":{"date-parts":[[2026,8,13]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                <jats:p>Single-cell atlases of the tumor-immune microenvironment have defined numerous fine-grained immune cell states, but each study uses its own nomenclature and procedure for annotating cell types. Transferring annotations from a reference atlas to a query dataset is complicated by both batch effects and by the presence of query populations that the reference does not contain. Here we present scDIVA, a semi-supervised deep generative model that adapts the Domain Invariant Variational Autoencoder to scRNA-seq for fine-grained tumor-immune label transfer. Three encoders disentangle each cell\u2019s expression profile into separate latent subspaces for cell type, batch, and residual variation; a single decoder reconstructs the cell\u2019s expression profile from all three latent embeddings, and auxiliary classifiers on the cell type and batch embeddings encourage each encoder to capture the respective source of variation; scDIVA\u2019s cell type embeddings are batch-invariant by construction rather than through explicit or adversarial correction. Benchmarked against four established reference-mapping approaches\u2014Harmony\/Symphony, scANVI with scArches, scPoli with scArches, and Seurat with label transfer\u2014across six tumor-immune atlases spanning five cancer types, scDIVA achieved the highest mean macro-F1 in five of six atlases and the highest biological conservation scores. To detect query-enriched populations, we adapted the Milo differential abundance (DA) framework, added a directional test, corrected spatialFDR weighting, and parallelized neighborhood distances for atlas-scale data, and applied it to scDIVA\u2019s cell type embeddings with reference- versus-query membership as the condition. This design correctly identified cell types held out from the reference as OOR and flagged exhausted CD8 T cells from a tumor-immune atlas as OOR relative to a healthy pan-tissue immune reference; conversely, this procedure confirmed a conserved immune landscape between two independent colorectal cancer cohorts. scDIVA thus couples fine-grained annotation and integration with an FDR-controlled test for states the reference lacks, solving both tasks required for accurate label transfer in tumor-immune scRNA-seq atlases.<\/jats:p>","DOI":"10.64898\/2026.08.07.743520","type":"posted-content","created":{"date-parts":[[2026,8,14]],"date-time":"2026-08-14T01:45:20Z","timestamp":1786671920000},"source":"Crossref","is-referenced-by-count":0,"title":["scDIVA: semi-supervised integration and fine-grained annotation of tumor-immune single-cell atlases"],"prefix":"10.64898","author":[{"ORCID":"https:\/\/orcid.org\/0009-0008-7089-435X","authenticated-orcid":false,"given":"Viraj","family":"Rapolu","sequence":"first","affiliation":[{"name":"Computational and Systems Biology Program, Memorial Sloan Kettering Cancer Center, New York, NY 10065"},{"name":"Tri-Institutional Program in Computational Biology and Medicine, New York, NY 10021"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-1723-9355","authenticated-orcid":false,"given":"Alireza","family":"Karbalayghareh","sequence":"additional","affiliation":[{"name":"Computational and Systems Biology Program, Memorial Sloan Kettering Cancer Center, New York, NY 10065"},{"name":"Chan Zuckerberg Biohub, New York, NY 10027"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Brennan","family":"Lee","sequence":"additional","affiliation":[{"name":"Computational and Systems Biology Program, Memorial Sloan Kettering Cancer Center, New York, NY 10065"},{"name":"Bristol Myers Squibb, Cambridge, MA 02141"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-1363-5235","authenticated-orcid":false,"given":"Wilfred","family":"Wong","sequence":"additional","affiliation":[{"name":"Computational and Systems Biology Program, Memorial Sloan Kettering Cancer Center, New York, NY 10065"},{"name":"Tri-Institutional Program in Computational Biology and Medicine, New York, NY 10021"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-4571-5910","authenticated-orcid":false,"given":"Christina S.","family":"Leslie","sequence":"additional","affiliation":[{"name":"Computational and Systems Biology Program, Memorial Sloan Kettering Cancer Center, New York, NY 10065"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"54368","reference":[{"key":"2026081712000974000_2026.08.07.743520v1.1","doi-asserted-by":"publisher","DOI":"10.1093\/nar\/gkae1142"},{"key":"2026081712000974000_2026.08.07.743520v1.2","doi-asserted-by":"publisher","DOI":"10.1126\/science.abl5197"},{"key":"2026081712000974000_2026.08.07.743520v1.3","doi-asserted-by":"crossref","first-page":"1088","DOI":"10.1038\/s43018-025-00957-8","article-title":"The Curated Cancer Cell Atlas provides a comprehensive characterization of tumors at single-cell resolution","volume":"6","year":"2025","journal-title":"Nature Cancer"},{"key":"2026081712000974000_2026.08.07.743520v1.4","doi-asserted-by":"publisher","DOI":"10.1016\/j.cell.2018.10.038"},{"key":"2026081712000974000_2026.08.07.743520v1.5","doi-asserted-by":"publisher","DOI":"10.1038\/s41568-019-0235-4"},{"key":"2026081712000974000_2026.08.07.743520v1.6","doi-asserted-by":"publisher","DOI":"10.1016\/j.ccell.2023.03.009"},{"key":"2026081712000974000_2026.08.07.743520v1.7","doi-asserted-by":"publisher","DOI":"10.1038\/s41591-023-02371-y"},{"key":"2026081712000974000_2026.08.07.743520v1.8","doi-asserted-by":"crossref","first-page":"145","DOI":"10.1016\/j.immuni.2025.11.014","article-title":"Opposing functions of distinct regulatory T cell subsets in colorectal cancer","volume":"59","year":"2026","journal-title":"Immunity"},{"key":"2026081712000974000_2026.08.07.743520v1.9","doi-asserted-by":"publisher","DOI":"10.7554\/eLife.27041"},{"key":"2026081712000974000_2026.08.07.743520v1.10","doi-asserted-by":"crossref","first-page":"159","DOI":"10.1186\/s13059-026-04055-5","article-title":"Benchmarking single-cell tumor immune atlases and application for uncovering cell states related to immunotherapy response","volume":"27","year":"2026","journal-title":"Genome Biol"},{"key":"2026081712000974000_2026.08.07.743520v1.11","first-page":"1","article-title":"Toward informed batch correction for single-cell transcriptome integration","volume":"6","year":"2026","journal-title":"Nat. Comput. Sci"},{"key":"2026081712000974000_2026.08.07.743520v1.12","doi-asserted-by":"publisher","DOI":"10.1038\/s41592-019-0619-0"},{"key":"2026081712000974000_2026.08.07.743520v1.13","doi-asserted-by":"publisher","DOI":"10.1038\/s41467-021-25957-x"},{"key":"2026081712000974000_2026.08.07.743520v1.14","doi-asserted-by":"publisher","DOI":"10.1016\/j.cell.2019.05.031"},{"key":"2026081712000974000_2026.08.07.743520v1.15","doi-asserted-by":"publisher","DOI":"10.1038\/s41592-018-0229-21"},{"key":"2026081712000974000_2026.08.07.743520v1.16","doi-asserted-by":"publisher","DOI":"10.15252\/msb.20209620"},{"key":"2026081712000974000_2026.08.07.743520v1.17","doi-asserted-by":"publisher","DOI":"10.1038\/s41587-021-01001-7"},{"key":"2026081712000974000_2026.08.07.743520v1.18","doi-asserted-by":"publisher","DOI":"10.1038\/s41592-023-02035-2"},{"key":"2026081712000974000_2026.08.07.743520v1.19","doi-asserted-by":"crossref","first-page":"lqad070","DOI":"10.1093\/nargab\/lqad070","article-title":"Single-cell reference mapping to construct and extend cell-type hierarchies","volume":"5","year":"2023","journal-title":"NAR Genomics Bioinforma"},{"key":"2026081712000974000_2026.08.07.743520v1.20","doi-asserted-by":"publisher","DOI":"10.1038\/s41467-023-43590-8"},{"key":"2026081712000974000_2026.08.07.743520v1.21","doi-asserted-by":"publisher","DOI":"10.1038\/s41587-023-02079-x"},{"key":"2026081712000974000_2026.08.07.743520v1.22","doi-asserted-by":"publisher","DOI":"10.1101\/2024.11.06.622266"},{"key":"2026081712000974000_2026.08.07.743520v1.23","doi-asserted-by":"publisher","DOI":"10.1101\/2024.12.06.627196"},{"key":"2026081712000974000_2026.08.07.743520v1.24","doi-asserted-by":"publisher","DOI":"10.1101\/2025.11.03.683998"},{"key":"2026081712000974000_2026.08.07.743520v1.25","doi-asserted-by":"publisher","DOI":"10.1038\/s41588-023-01523-7"},{"key":"2026081712000974000_2026.08.07.743520v1.26","doi-asserted-by":"publisher","DOI":"10.1038\/s41587-021-01033-z"},{"key":"2026081712000974000_2026.08.07.743520v1.27","doi-asserted-by":"publisher","DOI":"10.1038\/s41592-023-01955-3"},{"key":"2026081712000974000_2026.08.07.743520v1.28","doi-asserted-by":"crossref","first-page":"852","DOI":"10.1038\/s42256-022-00534-z","article-title":"scBERT as a large-scale pretrained deep language model for cell type annotation of single-cell RNA-seq data","volume":"4","year":"2022","journal-title":"Nat. Mach. Intell"},{"key":"2026081712000974000_2026.08.07.743520v1.29","unstructured":"Transfer learning enables predictions in network biology | Nature. https:\/\/www.nature.com\/articles\/s41586-023-06139-9."},{"key":"2026081712000974000_2026.08.07.743520v1.30","unstructured":"scGPT: toward building a foundation model for single-cell multi-omics using generative AI | Nature Methods. https:\/\/www.nature.com\/articles\/s41592-024-02201-0."},{"key":"2026081712000974000_2026.08.07.743520v1.31","doi-asserted-by":"publisher","DOI":"10.1038\/s41592-024-02305-7"},{"key":"2026081712000974000_2026.08.07.743520v1.32","doi-asserted-by":"publisher","DOI":"10.1038\/s41586-026-10689-z"},{"key":"2026081712000974000_2026.08.07.743520v1.33","unstructured":"Mitra, S. , Li, J. & Leslie, C. S. New frontiers for AI in single-cell genomics. (2026)."},{"key":"2026081712000974000_2026.08.07.743520v1.34","doi-asserted-by":"crossref","first-page":"1085","DOI":"10.1038\/s41586-024-08411-y","article-title":"A cell atlas foundation model for scalable search of similar human cells","volume":"638","year":"2025","journal-title":"Nature"},{"key":"2026081712000974000_2026.08.07.743520v1.35","doi-asserted-by":"crossref","first-page":"1447","DOI":"10.1038\/s41592-026-03120-y","article-title":"Evaluating the role of pretraining dataset size and diversity on single-cell foundation model performance","volume":"23","year":"2026","journal-title":"Nature Methods"},{"key":"2026081712000974000_2026.08.07.743520v1.36","doi-asserted-by":"publisher","DOI":"10.64898\/2026.03.10.710748"},{"key":"2026081712000974000_2026.08.07.743520v1.37","doi-asserted-by":"publisher","DOI":"10.64898\/2026.04.17.719314"},{"key":"2026081712000974000_2026.08.07.743520v1.38","unstructured":"Ilse, M. , Tomczak, J. M. , Louizos, C. & Welling, M. DIVA: Domain Invariant Variational Autoencoders. in Proceedings of the Third Conference on Medical Imaging with Deep Learning 322\u2013348 (PMLR, 2020)."},{"key":"2026081712000974000_2026.08.07.743520v1.39","doi-asserted-by":"publisher","DOI":"10.1016\/j.ccell.2021.03.007"},{"key":"2026081712000974000_2026.08.07.743520v1.40","doi-asserted-by":"publisher","DOI":"10.1016\/j.cell.2021.08.003"},{"key":"2026081712000974000_2026.08.07.743520v1.41","doi-asserted-by":"publisher","DOI":"10.1016\/j.ccell.2021.09.008"},{"key":"2026081712000974000_2026.08.07.743520v1.42","doi-asserted-by":"publisher","DOI":"10.1016\/j.ccell.2021.10.009"},{"key":"2026081712000974000_2026.08.07.743520v1.43","doi-asserted-by":"publisher","DOI":"10.1038\/s41586-022-05496-1"},{"key":"2026081712000974000_2026.08.07.743520v1.44","doi-asserted-by":"publisher","DOI":"10.1016\/j.cell.2020.03.048"},{"key":"2026081712000974000_2026.08.07.743520v1.45","doi-asserted-by":"publisher","DOI":"10.1038\/s41592-021-01336-8"},{"key":"2026081712000974000_2026.08.07.743520v1.46","doi-asserted-by":"publisher","DOI":"10.1038\/s41590-025-02347-9"},{"key":"2026081712000974000_2026.08.07.743520v1.47","doi-asserted-by":"crossref","first-page":"110","DOI":"10.1038\/s41590-025-02352-y","article-title":"Tissue-resident exhausted and memory CD8+ T cells have distinct ontogeny, function and roles in disease","volume":"27","year":"2026","journal-title":"Nat. Immunol"},{"key":"2026081712000974000_2026.08.07.743520v1.48","doi-asserted-by":"publisher","DOI":"10.64898\/2026.05.30.728980"}],"container-title":[],"original-title":[],"link":[{"URL":"https:\/\/syndication.highwire.org\/content\/doi\/10.64898\/2026.08.07.743520","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,8,17]],"date-time":"2026-08-17T19:00:18Z","timestamp":1786993218000},"score":1,"resource":{"primary":{"URL":"http:\/\/biorxiv.org\/lookup\/doi\/10.64898\/2026.08.07.743520"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,8,13]]},"references-count":48,"URL":"https:\/\/doi.org\/10.64898\/2026.08.07.743520","relation":{},"subject":[],"published":{"date-parts":[[2026,8,13]]},"subtype":"preprint"}}