{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,9,29]],"date-time":"2026-09-29T06:04:42Z","timestamp":1790661882598,"version":"4.1.0"},"reference-count":35,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2026,9,29]],"date-time":"2026-09-29T00:00:00Z","timestamp":1790640000000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"DOI":"10.13039\/501100000155","name":"Social Sciences and Humanities Research Council of Canada","doi-asserted-by":"publisher","award":["NFRFE-2024-00669"],"award-info":[{"award-number":["NFRFE-2024-00669"]}],"id":[{"id":"10.13039\/501100000155","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Bioinform."],"abstract":"<jats:sec>\n                    <jats:title>Background<\/jats:title>\n                    <jats:p>Breast, endometrial, and cervical cancers share a common challenge: molecular heterogeneity that single-modality profiling cannot fully resolve. Gene expression and DNA methylation capture distinct and partially non-overlapping regulatory programs, yet most subtyping frameworks treat them independently or merge them without accounting for modality-specific noise and scale.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Objective<\/jats:title>\n                    <jats:p>To develop an interpretable multimodal framework that integrates transcriptomic and epigenomic signals for principled molecular subtype discovery and molecular patient stratification with exploratory assessment of outcome associations across women\u2019s cancers.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Methods<\/jats:title>\n                    <jats:p>\n                      We developed STRIVE (Stratification via Integrated Variational Embeddings), an attention-based joint variational autoencoder that learns modality-aware embeddings, fuses them into a shared latent space, and supports\n                      <jats:italic>post hoc<\/jats:italic>\n                      feature attribution. Subtypes were defined by K-means clustering on latent embeddings, cluster-defining signatures were derived using standardized mean differences (q &amp;lt; 0.05), and within-subtype survival prediction was assessed via permutation feature importance across multiple classifiers.\n                    <\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>In TCGA-BRCA (n = 866), STRIVE identified six molecular subtypes with coherent multimodal signatures, with DNA-methylation features prominently represented in specific cluster signatures. Within-subtype survival-status prediction showed variable performance across classifiers and subtypes. Replication in TCGA-UCEC (n = 454) and TCGA-CESC (n = 296) recapitulated key signature patterns, with DNA-methylation features prominently represented in specific cluster signatures and more consistent predictive performance in UCEC than CESC.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Conclusion<\/jats:title>\n                    <jats:p>STRIVE demonstrates that attention-weighted multimodal integration surfaces molecularly coherent subtypes in which cluster-defining signatures frequently included epigenomic features, a structure that single-modality approaches would obscure. These findings position STRIVE as a framework for multimodal molecular stratification and future evaluation of subtype-specific clinical associations.<\/jats:p>\n                  <\/jats:sec>","DOI":"10.3389\/fbinf.2026.1901464","type":"journal-article","created":{"date-parts":[[2026,9,29]],"date-time":"2026-09-29T05:50:57Z","timestamp":1790661057000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":0,"title":["STRIVE: interpretable multimodal subtype discovery in Women\u2019s cancers using an attention-weighted joint variational autoencoder"],"prefix":"10.3389","volume":"6","author":[{"given":"Zunaira","family":"Mehmood","sequence":"first","affiliation":[{"name":"Department of Biostatistics, Dalla Lana School of Public Health, University of Toronto","place":["Toronto, ON, Canada"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Divya","family":"Sharma","sequence":"additional","affiliation":[{"name":"Department of Mathematics and Statistics at York University","place":["North York, ON, Canada"]},{"name":"Department of Biostatistics, University Health Network","place":["Toronto, ON, Canada"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1965","published-online":{"date-parts":[[2026,9,29]]},"reference":[{"key":"B1","doi-asserted-by":"publisher","first-page":"216","DOI":"10.1186\/s13059-025-03684-6","article-title":"FedscGen: privacy-preserving federated batch effect correction of single-cell RNA sequencing data","volume":"26","author":"Bakhtiari","year":"2025","journal-title":"Genome Biol."},{"key":"B2","doi-asserted-by":"publisher","first-page":"95","DOI":"10.1016\/j.cels.2020.05.010","article-title":"Solo: doublet identification in single-cell RNA-seq via semi-supervised deep learning","volume":"11","author":"Bernstein","year":"2020","journal-title":"Cell Syst."},{"key":"B3","doi-asserted-by":"publisher","first-page":"4151","DOI":"10.1038\/s41467-025-59499-3","article-title":"Design of diverse, functional mitochondrial targeting sequences across eukaryotic organisms using variational autoencoder","volume":"16","author":"Boob","year":"2025","journal-title":"Nat. Commun."},{"key":"B4","doi-asserted-by":"publisher","first-page":"7419","DOI":"10.1038\/s41467-022-35094-8","article-title":"A unified computational framework for single-cell data integration with optimal transport","volume":"13","author":"Cao","year":"2022","journal-title":"Nat. Commun."},{"key":"B5","doi-asserted-by":"publisher","first-page":"2973","DOI":"10.1038\/s41467-024-47418-x","article-title":"scButterfly: a versatile single-cell cross-modality translation method via dual-aligned variational autoencoders","volume":"15","author":"Cao","year":"2024","journal-title":"Nat. Commun."},{"key":"B6","doi-asserted-by":"publisher","first-page":"29","DOI":"10.1186\/s13059-023-02850-y","article-title":"siVAE: interpretable deep generative models for single-cell transcriptomes","volume":"24","author":"Choi","year":"2023","journal-title":"Genome Biol."},{"key":"B7","doi-asserted-by":"publisher","first-page":"188","DOI":"10.1186\/s13059-021-02407-x","article-title":"Optimizing expression quantitative trait locus mapping workflows for single-cell studies","volume":"22","author":"Cuomo","year":"2021","journal-title":"Genome Biol."},{"key":"B8","doi-asserted-by":"publisher","first-page":"932","DOI":"10.1016\/j.jbi.2010.07.001","article-title":"Independent component analysis: mining microarray data for fundamental human gene expression modules","volume":"43","author":"Engreitz","year":"2010","journal-title":"J. Biomed. Inf."},{"key":"B9","doi-asserted-by":"publisher","first-page":"115","DOI":"10.1038\/ng.3173","article-title":"Gene expression analysis identifies global gene dosage sensitivity in cancer","volume":"47","author":"Fehrmann","year":"2015","journal-title":"Nat. Genet."},{"key":"B10","doi-asserted-by":"publisher","first-page":"2211","DOI":"10.1093\/bioinformatics\/btt351","article-title":"CellMix: a comprehensive toolbox for gene expression deconvolution","volume":"29","author":"Gaujoux","year":"2013","journal-title":"Bioinformatics"},{"key":"B11","doi-asserted-by":"publisher","first-page":"675","DOI":"10.1038\/s41587-020-0546-8","article-title":"Visualizing and interpreting cancer genomics data via the xena platform","volume":"38","author":"Goldman","year":"2020","journal-title":"Nat. Biotechnol."},{"key":"B12","doi-asserted-by":"publisher","first-page":"351","DOI":"10.1186\/s13059-021-02556-z","article-title":"Cobolt: integrative analysis of multimodal single-cell sequencing data","volume":"22","author":"Gong","year":"2021","journal-title":"Genome Biol."},{"key":"B13","doi-asserted-by":"publisher","first-page":"125","DOI":"10.1186\/s13059-023-02954-5","article-title":"SeATAC: a tool for exploring the chromatin landscape and the role of pioneer factors","volume":"24","author":"Gong","year":"2023","journal-title":"Genome Biol."},{"key":"B15","doi-asserted-by":"publisher","first-page":"8","DOI":"10.1186\/s13059-021-02593-8","article-title":"scDALI: modeling allelic heterogeneity in single cells reveals context-specific genetic regulation","volume":"23","author":"Heinen","year":"2022","journal-title":"Genome Biol."},{"key":"B17","doi-asserted-by":"publisher","first-page":"149","DOI":"10.1186\/s13059-020-02055-7","article-title":"Enhanced integrated gradients: improving interpretability of deep learning models using splicing codes as a case study","volume":"21","author":"Jha","year":"2020","journal-title":"Genome Biol."},{"key":"B18","doi-asserted-by":"publisher","first-page":"309","DOI":"10.1186\/s13059-024-03431-3","article-title":"SyntheVAEiser: augmenting traditional machine learning methods with VAE-based gene expression sample generation for improved cancer subtype predictions","volume":"25","author":"Karlberg","year":"2024","journal-title":"Genome Biol."},{"key":"B19","doi-asserted-by":"publisher","first-page":"501","DOI":"10.2144\/000112950","article-title":"A review of independent component analysis application to microarray gene expression data","volume":"45","author":"Kong","year":"2008","journal-title":"BioTechniques"},{"key":"B21","doi-asserted-by":"publisher","first-page":"174","DOI":"10.1186\/s13059-022-02723-w","article-title":"BindVAE: dirichlet Variational Autoencoders for de novo Motif Discovery from Accessible Chromatin","volume":"23","author":"Kshirsagar","year":"2022","journal-title":"Genome Biol."},{"key":"B22","doi-asserted-by":"publisher","first-page":"32173","DOI":"10.1038\/s41598-025-12021-7","article-title":"Benchmarking of dimensionality reduction methods to capture drug response in transcriptome data","volume":"15","author":"Kwon","year":"2025","journal-title":"Sci. Rep."},{"key":"B23","doi-asserted-by":"publisher","first-page":"7705","DOI":"10.1038\/s41467-022-35031-9","article-title":"Clustering of single-cell multi-omics data with a multimodal deep learning method","volume":"13","author":"Lin","year":"2022","journal-title":"Nat. Commun."},{"key":"B24","doi-asserted-by":"publisher","first-page":"49","DOI":"10.1016\/j.cels.2020.05.007","article-title":"A generative neural network for maximizing fitness and diversity of synthetic DNA and protein sequences","volume":"11","author":"Linder","year":"2020","journal-title":"Cell Syst."},{"key":"B25","doi-asserted-by":"publisher","first-page":"3632","DOI":"10.1038\/s41596-020-0391-8","article-title":"Jointly defining cell types from multiple single-cell datasets using LIGER","volume":"15","author":"Liu","year":"2020","journal-title":"Nat. Protoc."},{"key":"B26","doi-asserted-by":"publisher","first-page":"411","DOI":"10.1016\/j.cels.2024.04.004","article-title":"Reconstructing developmental trajectories using latent dynamical systems and time-resolved transcriptomics","volume":"15","author":"Maizels","year":"2024","journal-title":"Cell Syst."},{"key":"B27","doi-asserted-by":"publisher","first-page":"2326","DOI":"10.1038\/s41467-022-29843-y","article-title":"A deep Siamese neural network improves metagenome-assembled genomes in microbiome datasets across different environments","volume":"13","author":"Pan","year":"2022","journal-title":"Nat. Commun."},{"key":"B29","doi-asserted-by":"publisher","first-page":"844","DOI":"10.1016\/j.ard.2025.01.012","article-title":"Deep learning\u2013based clustering for endotyping and post-arthroplasty response classification using knee osteoarthritis multiomic data","volume":"84","author":"Rockel","year":"2025","journal-title":"Ann. Rheumatic Dis."},{"key":"B30","doi-asserted-by":"publisher","first-page":"64","DOI":"10.1186\/s13059-025-03519-4","article-title":"scVAEDer: integrating deep diffusion models and variational autoencoders for single-cell transcriptomics analysis","volume":"26","author":"Sadria","year":"2025","journal-title":"Genome Biol."},{"key":"B31","doi-asserted-by":"publisher","first-page":"571","DOI":"10.1016\/j.coi.2013.09.015","article-title":"Computational deconvolution: extracting cell type-specific information from heterogeneous samples","volume":"25","author":"Shen-Orr","year":"2013","journal-title":"Curr. Opin. Immunol."},{"key":"B34","doi-asserted-by":"publisher","first-page":"107220","DOI":"10.1016\/j.compbiomed.2023.107220","article-title":"The prediction of drug sensitivity by multi-omics fusion reveals the heterogeneity of drug response in pan-cancer","volume":"163","author":"Wang","year":"2023","journal-title":"Comput. Biol. Med."},{"key":"B35","doi-asserted-by":"publisher","first-page":"149","DOI":"10.1186\/s13059-024-03293-9","article-title":"TMO-Net: an explainable pretrained multi-omics model for multi-task learning in oncology","volume":"25","author":"Wang","year":"2024","journal-title":"Genome Biol."},{"key":"B36","doi-asserted-by":"publisher","first-page":"109","DOI":"10.1186\/s13059-020-02021-3","article-title":"Compressing gene expression data using multiple latent space dimensionalities learns complementary biological representations","volume":"21","author":"Way","year":"2020","journal-title":"Genome Biol."},{"key":"B38","doi-asserted-by":"publisher","first-page":"7603","DOI":"10.1038\/s41467-023-43220-3","article-title":"SPACEL: deep learning-based characterization of spatial transcriptome architectures","volume":"14","author":"Xu","year":"2023","journal-title":"Nat. Commun."},{"key":"B39","doi-asserted-by":"publisher","first-page":"265","DOI":"10.1186\/s13059-025-03719-y","article-title":"Comparative benchmarking of single-cell clustering algorithms for transcriptomic and proteomic data","volume":"26","author":"Yin","year":"2025","journal-title":"Genome Biol."},{"key":"B40","doi-asserted-by":"publisher","first-page":"1226","DOI":"10.1038\/s41467-021-21254-9","article-title":"Uniform genomic data analysis in the NCI genomic data commons","volume":"12","author":"Zhang","year":"2021","journal-title":"Nat. Commun."},{"key":"B41","doi-asserted-by":"publisher","first-page":"222","DOI":"10.1186\/s13059-022-02774-z","article-title":"Normalization and De-Noising of single-cell Hi-C data with BandNorm and scVI-3D","volume":"23","author":"Zheng","year":"2022","journal-title":"Genome Biol."},{"key":"B42","doi-asserted-by":"publisher","first-page":"1562","DOI":"10.1016\/j.ajhg.2025.05.015","article-title":"Applying multimodal AI to physiological waveforms improves genetic prediction of cardiovascular traits","volume":"112","author":"Zhou","year":"2025","journal-title":"Am. J. Hum. Genet."}],"container-title":["Frontiers in Bioinformatics"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fbinf.2026.1901464\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,9,29]],"date-time":"2026-09-29T05:50:58Z","timestamp":1790661058000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fbinf.2026.1901464\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,9,29]]},"references-count":35,"alternative-id":["10.3389\/fbinf.2026.1901464"],"URL":"https:\/\/doi.org\/10.3389\/fbinf.2026.1901464","relation":{},"ISSN":["2673-7647"],"issn-type":[{"value":"2673-7647","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,9,29]]},"article-number":"1901464"}}