{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,1]],"date-time":"2026-10-01T16:58:16Z","timestamp":1790873896238,"version":"4.1.0"},"reference-count":82,"publisher":"Cold Spring Harbor Laboratory","issue":"10","content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["Genome Res."],"accepted":{"date-parts":[[2026,8,11]]},"published-print":{"date-parts":[[2026,10]]},"abstract":"<jats:p>Multimodal single-cell assays profile complementary layers of cell state, but integration is complicated by modality mismatch, sparsity, and uneven cohort coverage. Here, we present Unified Variational Inference (UniVI), a scalable mixture-of-experts \u03b2-variational autoencoder that learns a shared latent space while preserving modality-specific structure. UniVI couples modality-specific encoders\/decoders with a shared latent prior and a symmetric cross-modal alignment objective, enabling consistent integration of paired measurements without curated feature-link graphs or preannotated reference atlases; optional supervised heads can be added when labels are available. Across paired RNA\u2013protein (CITE-seq) and RNA\u2013chromatin (10x Genomics Multiome, SHARE-seq) data spanning human PBMCs and mouse back skin\u2014a nonhematopoietic tissue with continuous differentiation hierarchies\u2014UniVI produces coherent embeddings, improves label transfer, and enables cross-modal reconstruction and denoising. Extending to trimodal measurements, UniVI maintains robust three-way alignment among RNA, chromatin accessibility, and surface proteins (TEA-seq), and accommodates DNA methylation in a paired scNMT-seq mouse gastrulation proof-of-concept under beta-binomial likelihoods. Performance degrades gracefully under severe cell type imbalance and in the presence of modality-exclusive populations. In an acute myeloid leukemia mosaic design, a paired RNA\u2013protein bridge anchors independent RNA-only and protein+genotype cohorts, revealing genotype-associated neighborhoods that sharpen with mutation-aware fine-tuning. UniVI thus provides a flexible, interpretable framework for multimodal integration across paired, trimodal, and mosaic study designs and supports practical reference-to-query projection in partially observed studies.<\/jats:p>","DOI":"10.1101\/gr.281431.125","type":"journal-article","created":{"date-parts":[[2026,9,16]],"date-time":"2026-09-16T19:40:24Z","timestamp":1789587624000},"page":"2068-2090","source":"Crossref","is-referenced-by-count":0,"title":["Unifying multimodal single-cell data with a mixture-of-experts \u03b2-variational autoencoder framework"],"prefix":"10.1101","volume":"36","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-1234-2118","authenticated-orcid":false,"given":"Andrew J.","family":"Ashford","sequence":"first","affiliation":[{"name":"Department of Molecular and Medical Genetics, Oregon Health & Science University, Portland, Oregon 97239, USA;"},{"name":"Computational Biology Department, Oregon Health & Science University, Portland, Oregon 97201, USA;"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0006-4367-2386","authenticated-orcid":false,"given":"Trevor","family":"Enright","sequence":"additional","affiliation":[{"name":"Computational Biology Department, Oregon Health & Science University, Portland, Oregon 97201, USA;"},{"name":"Cancer Early Detection Advanced Research Center (CEDAR), Knight Cancer Institute, Oregon Health & Science University, Portland, Oregon 97201, USA;"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-9489-2246","authenticated-orcid":false,"given":"Julia","family":"Somers","sequence":"additional","affiliation":[{"name":"Cancer Early Detection Advanced Research Center (CEDAR), Knight Cancer Institute, Oregon Health & Science University, Portland, Oregon 97201, USA;"},{"name":"Biomedical Engineering Department, Oregon Health & Science University, Portland, Oregon 97201, USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-8105-2440","authenticated-orcid":false,"given":"Olga","family":"Nikolova","sequence":"additional","affiliation":[{"name":"Cancer Early Detection Advanced Research Center (CEDAR), Knight Cancer Institute, Oregon Health & Science University, Portland, Oregon 97201, USA;"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-3663-7113","authenticated-orcid":false,"given":"Emek","family":"Demir","sequence":"additional","affiliation":[{"name":"Department of Molecular and Medical Genetics, Oregon Health & Science University, Portland, Oregon 97239, USA;"},{"name":"Computational Biology Department, Oregon Health & Science University, Portland, Oregon 97201, USA;"},{"name":"Cancer Early Detection Advanced Research Center (CEDAR), Knight Cancer Institute, Oregon Health & Science University, Portland, Oregon 97201, USA;"},{"name":"Biomedical Engineering Department, Oregon Health & Science University, Portland, Oregon 97201, USA"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"246","published-online":{"date-parts":[[2026,9,16]]},"reference":[{"key":"2026100109051462000_36.10.2068.1","unstructured":"10x Genomics. 2021a. 10k human PBMCs, multiome v1.0, Chromium X, single-cell multiome ATAC + gene expression dataset generated by a Chromium X series instrument, analyzed using Cell Ranger ARC 2.0.0."},{"key":"2026100109051462000_36.10.2068.2","unstructured":"10x Genomics. 2021b. PBMCs from a healthy donor \u2013 no cell sorting (10k), single-cell multiome ATAC + gene expression dataset generated by a Chromium Controller instrument and analyzed by Cell Ranger ARC 2.0.0."},{"key":"2026100109051462000_36.10.2068.3","unstructured":"Andrew G , Arora R , Bilmes J , Livescu K . 2013. Deep canonical correlation analysis. In Proceedings of the 30th International Conference on Machine Learning (ed. Dasgupta S, McAllester D), Vol. 28, pp. 1247\u20131255. PMLR, Atlanta."},{"key":"2026100109051462000_36.10.2068.4","doi-asserted-by":"publisher","DOI":"10.1038\/nmeth.3728"},{"key":"2026100109051462000_36.10.2068.5","doi-asserted-by":"publisher","DOI":"10.15252\/msb.20178124"},{"key":"2026100109051462000_36.10.2068.6","doi-asserted-by":"publisher","DOI":"10.1038\/s41586-019-1825-8"},{"key":"2026100109051462000_36.10.2068.7","doi-asserted-by":"publisher","DOI":"10.1186\/s13059-020-02015-1"},{"key":"2026100109051462000_36.10.2068.8","doi-asserted-by":"publisher","DOI":"10.1038\/s41587-021-00895-7"},{"key":"2026100109051462000_36.10.2068.9","doi-asserted-by":"publisher","DOI":"10.1016\/j.crmeth.2022.100182"},{"key":"2026100109051462000_36.10.2068.10","doi-asserted-by":"publisher","DOI":"10.1038\/s41592-023-01909-9"},{"key":"2026100109051462000_36.10.2068.11","doi-asserted-by":"publisher","DOI":"10.1038\/s41580-023-00615-w"},{"key":"2026100109051462000_36.10.2068.12","doi-asserted-by":"publisher","DOI":"10.1002\/0471142727.mb2129s109"},{"key":"2026100109051462000_36.10.2068.13","doi-asserted-by":"publisher","DOI":"10.1038\/s41592-023-01969-x"},{"key":"2026100109051462000_36.10.2068.14","doi-asserted-by":"publisher","DOI":"10.1038\/nbt.4096"},{"key":"2026100109051462000_36.10.2068.15","doi-asserted-by":"publisher","DOI":"10.1038\/s41587-022-01284-4"},{"key":"2026100109051462000_36.10.2068.16","doi-asserted-by":"publisher","DOI":"10.1038\/s41467-022-35094-8"},{"key":"2026100109051462000_36.10.2068.17","doi-asserted-by":"publisher","DOI":"10.1038\/s41588-021-00927-7"},{"key":"2026100109051462000_36.10.2068.18","doi-asserted-by":"publisher","DOI":"10.1038\/s41587-023-01935-0"},{"key":"2026100109051462000_36.10.2068.19","doi-asserted-by":"publisher","DOI":"10.1126\/science.aab1601"},{"key":"2026100109051462000_36.10.2068.20","doi-asserted-by":"publisher","DOI":"10.1038\/s41467-021-21810-3"},{"key":"2026100109051462000_36.10.2068.21","doi-asserted-by":"publisher","DOI":"10.1089\/cmb.2021.0446"},{"key":"2026100109051462000_36.10.2068.22","doi-asserted-by":"publisher","DOI":"10.1038\/s41587-020-0465-8"},{"key":"2026100109051462000_36.10.2068.23","doi-asserted-by":"publisher","DOI":"10.1038\/s41467-018-07931-2"},{"key":"2026100109051462000_36.10.2068.24","doi-asserted-by":"publisher","DOI":"10.1038\/s41592-025-02737-9"},{"key":"2026100109051462000_36.10.2068.25","doi-asserted-by":"publisher","DOI":"10.1038\/s41587-021-00867-x"},{"key":"2026100109051462000_36.10.2068.26","doi-asserted-by":"publisher","DOI":"10.1038\/s41592-020-01050-x"},{"key":"2026100109051462000_36.10.2068.27","doi-asserted-by":"publisher","DOI":"10.1186\/s13059-021-02556-z"},{"key":"2026100109051462000_36.10.2068.28","doi-asserted-by":"publisher","DOI":"10.1038\/s41588-021-00790-6"},{"key":"2026100109051462000_36.10.2068.29","doi-asserted-by":"publisher","DOI":"10.1038\/nbt.4091"},{"key":"2026100109051462000_36.10.2068.30","doi-asserted-by":"publisher","DOI":"10.1016\/j.cell.2021.04.048"},{"key":"2026100109051462000_36.10.2068.31","doi-asserted-by":"publisher","DOI":"10.1038\/s41587-023-01767-y"},{"key":"2026100109051462000_36.10.2068.32","doi-asserted-by":"publisher","DOI":"10.1162\/0899766042321814"},{"key":"2026100109051462000_36.10.2068.33","doi-asserted-by":"publisher","DOI":"10.1038\/s41587-023-02040-y"},{"key":"2026100109051462000_36.10.2068.34","doi-asserted-by":"publisher","DOI":"10.1038\/s41576-023-00586-w"},{"key":"2026100109051462000_36.10.2068.35","unstructured":"Higgins I , Matthey L , Pal A , Burgess C , Glorot X , Botvinick M , Mohamed S , Lerchner A . 2017. beta-VAE: learning basic visual concepts with a constrained variational framework. In Proceedings of the 5th International Conference on Learning Representations (ICLR), Toulon, France."},{"key":"2026100109051462000_36.10.2068.36","doi-asserted-by":"publisher","DOI":"10.1093\/biomet\/28.3-4.321"},{"key":"2026100109051462000_36.10.2068.37","doi-asserted-by":"publisher","DOI":"10.1007\/BF01908075"},{"key":"2026100109051462000_36.10.2068.38","doi-asserted-by":"publisher","DOI":"10.1093\/bioinformatics\/btac084"},{"key":"2026100109051462000_36.10.2068.39","doi-asserted-by":"publisher","DOI":"10.1186\/s13059-021-02565-y"},{"key":"2026100109051462000_36.10.2068.40","doi-asserted-by":"publisher","DOI":"10.1162\/neco.1994.6.2.181"},{"key":"2026100109051462000_36.10.2068.41","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.1312.6114"},{"key":"2026100109051462000_36.10.2068.42","doi-asserted-by":"publisher","DOI":"10.1038\/s43018-023-00656-2"},{"key":"2026100109051462000_36.10.2068.43","doi-asserted-by":"publisher","DOI":"10.1038\/s41592-019-0619-0"},{"key":"2026100109051462000_36.10.2068.44","doi-asserted-by":"publisher","DOI":"10.3390\/e19110631"},{"key":"2026100109051462000_36.10.2068.45","doi-asserted-by":"publisher","DOI":"10.1038\/s41587-021-01161-6"},{"key":"2026100109051462000_36.10.2068.46","doi-asserted-by":"publisher","DOI":"10.1038\/s41592-025-02856-3"},{"key":"2026100109051462000_36.10.2068.47","doi-asserted-by":"publisher","DOI":"10.1038\/s41592-018-0229-2"},{"key":"2026100109051462000_36.10.2068.48","unstructured":"Loshchilov I , Hutter F . 2019. Decoupled weight decay regularization. In International Conference on Learning Representations (ICLR), New Orleans."},{"key":"2026100109051462000_36.10.2068.49","doi-asserted-by":"publisher","DOI":"10.1038\/s41592-021-01336-8"},{"key":"2026100109051462000_36.10.2068.50","doi-asserted-by":"publisher","DOI":"10.1016\/j.cell.2020.09.056"},{"key":"2026100109051462000_36.10.2068.51","doi-asserted-by":"publisher","DOI":"10.21105\/joss.00861"},{"key":"2026100109051462000_36.10.2068.52","doi-asserted-by":"publisher","DOI":"10.1038\/s41592-019-0392-0"},{"key":"2026100109051462000_36.10.2068.53","doi-asserted-by":"publisher","DOI":"10.1038\/s41587-021-00927-2"},{"key":"2026100109051462000_36.10.2068.54","doi-asserted-by":"publisher","DOI":"10.1016\/j.crmeth.2021.100071"},{"key":"2026100109051462000_36.10.2068.55","doi-asserted-by":"publisher","DOI":"10.1038\/s41596-021-00573-7"},{"key":"2026100109051462000_36.10.2068.56","unstructured":"Muon Team. 2022. Processing chromatin accessibility of 10k PBMCs. Muon tutorials. https:\/\/muon-tutorials.readthedocs.io\/en\/latest\/single-cell-rna-atac\/pbmc10k\/2-Chromatin-Accessibility-Processing.html [accessed February 19, 2024]."},{"key":"2026100109051462000_36.10.2068.57","doi-asserted-by":"publisher","DOI":"10.1038\/nature20598"},{"key":"2026100109051462000_36.10.2068.58","unstructured":"Paszke A , Gross S , Massa F , Lerer A , Bradbury J , Chanan G , Killeen T , Lin Z , Gimelshein M , Antiga L , 2019. PyTorch: an imperative style, high-performance deep learning library. In Advances in Neural Information Processing Systems 32 (NeurIPS 2019), Vancouver (ed. Wallach H , ), pp. 8024\u20138035."},{"key":"2026100109051462000_36.10.2068.59","doi-asserted-by":"publisher","DOI":"10.1561\/2200000073"},{"key":"2026100109051462000_36.10.2068.60","doi-asserted-by":"publisher","DOI":"10.1038\/s41586-019-0933-9"},{"key":"2026100109051462000_36.10.2068.61","doi-asserted-by":"publisher","DOI":"10.1016\/j.molcel.2018.06.044"},{"key":"2026100109051462000_36.10.2068.62","doi-asserted-by":"publisher","DOI":"10.1016\/0377-0427(87)90125-7"},{"key":"2026100109051462000_36.10.2068.63","doi-asserted-by":"publisher","DOI":"10.1038\/s41587-019-0206-z"},{"key":"2026100109051462000_36.10.2068.64","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.1701.06538"},{"key":"2026100109051462000_36.10.2068.65","unstructured":"Shi Y , Siddharth N , Paige B , Torr PHS . 2019. Variational mixture-of-experts autoencoders for multi-modal deep generative models. In Advances in Neural Information Processing Systems 32 (NeurIPS 2019), Vancouver (ed. Wallach H , )."},{"key":"2026100109051462000_36.10.2068.66","doi-asserted-by":"crossref","unstructured":"Singh R , Demetci P , Bonora G , Ramani V , Lee C , Fang H , Duan Q . 2020. Unsupervised manifold alignment for single-cell multi-omics data. In Proceedings of the 11th ACM International Conference on Bioinformatics, Computational Biology and Health Informatics, Association for Computing Machinery, New York.","DOI":"10.1145\/3388440.3412410"},{"key":"2026100109051462000_36.10.2068.67","doi-asserted-by":"publisher","DOI":"10.1093\/bib\/bbab569"},{"key":"2026100109051462000_36.10.2068.68","doi-asserted-by":"publisher","DOI":"10.1093\/bioinformatics\/btaa843"},{"key":"2026100109051462000_36.10.2068.69","doi-asserted-by":"publisher","DOI":"10.1038\/nmeth.4380"},{"key":"2026100109051462000_36.10.2068.70","doi-asserted-by":"publisher","DOI":"10.1016\/j.cell.2019.05.031"},{"key":"2026100109051462000_36.10.2068.71","doi-asserted-by":"publisher","DOI":"10.1038\/s41592-021-01282-5"},{"key":"2026100109051462000_36.10.2068.72","doi-asserted-by":"publisher","DOI":"10.1038\/s41587-019-0379-5"},{"key":"2026100109051462000_36.10.2068.73","doi-asserted-by":"publisher","DOI":"10.7554\/eLife.63632"},{"key":"2026100109051462000_36.10.2068.74","doi-asserted-by":"publisher","DOI":"10.1038\/s41598-019-41695-z"},{"key":"2026100109051462000_36.10.2068.75","doi-asserted-by":"publisher","DOI":"10.1016\/j.cell.2019.01.031"},{"key":"2026100109051462000_36.10.2068.76","doi-asserted-by":"publisher","DOI":"10.21105\/joss.04371"},{"key":"2026100109051462000_36.10.2068.77","unstructured":"Wang C , Mahadevan S . 2009. A general framework for manifold alignment. In Proceedings of the AAAI Fall Symposium on Manifold Learning and its Applications, Arlington, VA, pp. 79\u201386. AAAI Press, Washington, DC."},{"key":"2026100109051462000_36.10.2068.78","doi-asserted-by":"publisher","DOI":"10.1016\/j.cell.2019.05.006"},{"key":"2026100109051462000_36.10.2068.79","doi-asserted-by":"publisher","DOI":"10.1186\/s13059-017-1382-0"},{"key":"2026100109051462000_36.10.2068.80","doi-asserted-by":"publisher","DOI":"10.1093\/bib\/bbae095"},{"key":"2026100109051462000_36.10.2068.81","doi-asserted-by":"publisher","DOI":"10.1038\/s41467-023-36066-2"},{"key":"2026100109051462000_36.10.2068.82","doi-asserted-by":"publisher","DOI":"10.1093\/bioinformatics\/btaf345"}],"container-title":["Genome Research"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/syndication.highwire.org\/content\/doi\/10.1101\/gr.281431.125","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,10,1]],"date-time":"2026-10-01T16:05:44Z","timestamp":1790870744000},"score":1,"resource":{"primary":{"URL":"http:\/\/genome.cshlp.org\/lookup\/doi\/10.1101\/gr.281431.125"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,9,16]]},"references-count":82,"journal-issue":{"issue":"10","published-online":{"date-parts":[[2026,10,1]]},"published-print":{"date-parts":[[2026,10]]}},"alternative-id":["10.1101\/gr.281431.125"],"URL":"https:\/\/doi.org\/10.1101\/gr.281431.125","relation":{},"ISSN":["1088-9051","1549-5469"],"issn-type":[{"value":"1088-9051","type":"print"},{"value":"1549-5469","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,9,16]]}}}