{"paper_id":"10229261-5de2-41e7-83d5-e85f74e4fac2","body_text":"Mutual complementarity maximization in multi-view proteomics enables robust biomarker prediction | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Mutual complementarity maximization in multi-view proteomics enables robust biomarker prediction Wilson Goh, hui Peng, Hexin Cai, Shifu Luo, Jinyan Li This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9367379/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted You are reading this latest preprint version Abstract Mass spectrometry (MS)-based proteomic quantification is a multi-dimensional problem. Current approaches focus only on single dimensions (views), which is a limited representation of the underlying proteome. Distinct quantification modes (“views”) encodes complementary biological information, but each view is separately explored and not well-integrated. This has led towards biased single-view studies in comparative proteomic analysis, including the identification of differentially expressed proteins (DEPs), which are important for drug target discovery and disease diagnosis. To exploit mutual multi-view complementarity, we introduce Multi-view Contrastive Proteomics (MCP), a semi-supervised representation learning framework that integrates these heterogeneous quantification views into a shared embedding space through contrastive learning. MCP’s novelty lies in the explicit modeling of cross-view interactions while mitigating label scarcity, technical noise, and class imbalance. By jointly leveraging complementary views rather than treating them independently, MCP maximizes biologically meaningful data for DEP detection. Benchmarked on 18 gold-standard spike-in datasets, MCP consistently predicted more true DEPs, achieving >20% increase in recall while maintaining precision (>90%) in comparison with prevailing statistical and machine learning baselines. MCP also provides interpretable insights into cross-view interactions, revealing deeper structural mechanisms underpinning proteomic data. When applied to clinical proteomics cohorts, MCP unlocks previously inaccessible biological signals: robustly identifying unique DEPs that enhance thyroid nodule stratification. MCP also identified VSIG4, a macrophage immune suppressive regulator, as a novel prognostic biomarker for colorectal cancer that was missed by conventional approaches; in perturbation studies, MCP recovers more validated IFN-γ-responsive proteins and pathways from both bulk and single-cell proteomics, confirming the roles by perturbation modeling and virtual cell generation. Overall, MCP establishes a general computational paradigm for complementarity maximization in multi-view proteomic data analysis, enabling deeper biological understanding and translational impact. MCP is available at http://www.ai4pro.tech:3838 . Biological sciences/Computational biology and bioinformatics/Computational models Biological sciences/Computational biology and bioinformatics/Proteome informatics Semi-supervised deep multi-view contrastive learning Proteomics Differentially expressed proteins computational biology Full Text Additional Declarations There is NO Competing Interest. Supplementary Files SupplementaryInformation.docx Supplementary Information for Mutual complementarity maximization in multi-view proteomics enables robust biomarker prediction SupplementaryData2.xlsx Supplementary Data 2 SupplementaryData3.xlsx Supplementary Data 3 SupplementaryData4.xlsx Supplementary Data 4 SupplementaryData8.xlsx Supplementary Data 8 SupplementaryData1.xlsx Supplementary Data 1 SupplementaryData5.xlsx Supplementary Data 5 SupplementaryData6.xlsx Supplementary Data 6 SupplementaryData7.xlsx Supplementary Data 7 Cite Share Download PDF Status: Under Review Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {\"props\":{\"pageProps\":{\"initialData\":{\"identity\":\"rs-9367379\",\"acceptedTermsAndConditions\":true,\"allowDirectSubmit\":false,\"archivedVersions\":[],\"articleType\":\"Article\",\"associatedPublications\":[],\"authors\":[{\"id\":624029768,\"identity\":\"298b0977-87d7-4580-afea-7ef9a8927ca6\",\"order_by\":0,\"name\":\"Wilson Goh\",\"email\":\"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA6ElEQVRIiWNgGAWjYDACZuYGMG3ADqEZGwhrgaox4DkAZhGhBabGQCKBSC0GxxmbP3zcUStnLvn2+GMeBhvZDQfYH7/Aq+UwY5vkzDPHjS1n5yU28zCkGW84wGNmgU+LGVALM2/bscQNt3MMgVoOJwK1sBkQ0NL8Gazl5hmQlv9ALezPCGlpkOZtq0nccIMHpOUAUAuD8QN8WuzBfmk7YGxwJsdw5hyDZOOZh3nM8OlgkOw/fPjDx7Y6OYPjZww+vKmwk+073v74A149EHAYSoM8wczAJkGEljoUHjMxtoyCUTAKRsHIAQDGh1AwOFa4JwAAAABJRU5ErkJggg==\",\"orcid\":\"https://orcid.org/0000-0003-3863-7501\",\"institution\":\"Nanyang Technological University\",\"correspondingAuthor\":true,\"prefix\":\"\",\"firstName\":\"Wilson\",\"middleName\":\"\",\"lastName\":\"Goh\",\"suffix\":\"\"},{\"id\":624029769,\"identity\":\"f07fddb0-c325-4c18-be20-4f6866ea1e80\",\"order_by\":1,\"name\":\"hui Peng\",\"email\":\"\",\"orcid\":\"https://orcid.org/0000-0002-5137-1827\",\"institution\":\"Shenzhen University of Advanced Technology\",\"correspondingAuthor\":false,\"prefix\":\"\",\"firstName\":\"hui\",\"middleName\":\"\",\"lastName\":\"Peng\",\"suffix\":\"\"},{\"id\":624029770,\"identity\":\"a946135d-3862-483d-b193-52ef13ebf432\",\"order_by\":2,\"name\":\"Hexin Cai\",\"email\":\"\",\"orcid\":\"\",\"institution\":\"Shenzhen University of Advanced Technology\",\"correspondingAuthor\":false,\"prefix\":\"\",\"firstName\":\"Hexin\",\"middleName\":\"\",\"lastName\":\"Cai\",\"suffix\":\"\"},{\"id\":624029771,\"identity\":\"81728c83-cea4-44d0-8482-e7c2adbceaf0\",\"order_by\":3,\"name\":\"Shifu Luo\",\"email\":\"\",\"orcid\":\"\",\"institution\":\"Shenzhen University of Advanced Technology\",\"correspondingAuthor\":false,\"prefix\":\"\",\"firstName\":\"Shifu\",\"middleName\":\"\",\"lastName\":\"Luo\",\"suffix\":\"\"},{\"id\":624029772,\"identity\":\"d8d9ccd3-2819-4ab2-822f-dab39acc82a2\",\"order_by\":4,\"name\":\"Jinyan Li\",\"email\":\"\",\"orcid\":\"\",\"institution\":\"Shenzhen University of Advanced Technology\",\"correspondingAuthor\":false,\"prefix\":\"\",\"firstName\":\"Jinyan\",\"middleName\":\"\",\"lastName\":\"Li\",\"suffix\":\"\"}],\"badges\":[],\"createdAt\":\"2026-04-09 10:46:35\",\"currentVersionCode\":1,\"declarations\":\"\",\"doi\":\"10.21203/rs.3.rs-9367379/v1\",\"doiUrl\":\"https://doi.org/10.21203/rs.3.rs-9367379/v1\",\"draftVersion\":[],\"editorialEvents\":[],\"editorialNote\":\"\",\"failedWorkflow\":false,\"files\":[{\"id\":108189838,\"identity\":\"c101d7e5-b9a6-40d7-86e2-af15f70bf350\",\"added_by\":\"auto\",\"created_at\":\"2026-04-30 09:43:13\",\"extension\":\"pdf\",\"order_by\":1,\"title\":\"\",\"display\":\"\",\"copyAsset\":false,\"role\":\"manuscript-pdf\",\"size\":2469324,\"visible\":true,\"origin\":\"\",\"legend\":\"\",\"description\":\"\",\"filename\":\"manuscript.pdf\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-9367379/v1_covered_bd34abe1-8ff4-4961-a038-ae9590bf9f65.pdf\"},{\"id\":108189662,\"identity\":\"08d8ab2c-93ef-4b29-81d7-0a08cdc4e952\",\"added_by\":\"auto\",\"created_at\":\"2026-04-30 09:42:33\",\"extension\":\"docx\",\"order_by\":1,\"title\":\"\",\"display\":\"\",\"copyAsset\":false,\"role\":\"supplement\",\"size\":4822727,\"visible\":true,\"origin\":\"\",\"legend\":\"Supplementary Information for Mutual complementarity maximization in multi-view proteomics enables robust biomarker prediction\",\"description\":\"\",\"filename\":\"SupplementaryInformation.docx\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-9367379/v1/99aa5b28598f3027773a53aa.docx\"},{\"id\":108189609,\"identity\":\"cbad745d-c7f5-4a67-99ec-0fac5c6b0ba2\",\"added_by\":\"auto\",\"created_at\":\"2026-04-30 09:42:24\",\"extension\":\"xlsx\",\"order_by\":2,\"title\":\"\",\"display\":\"\",\"copyAsset\":false,\"role\":\"supplement\",\"size\":216838,\"visible\":true,\"origin\":\"\",\"legend\":\"Supplementary Data 2\",\"description\":\"\",\"filename\":\"SupplementaryData2.xlsx\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-9367379/v1/d57e672ae0a12da03cf520d6.xlsx\"},{\"id\":108189611,\"identity\":\"74c3923e-4f48-43b9-84e2-af79c3223997\",\"added_by\":\"auto\",\"created_at\":\"2026-04-30 09:42:25\",\"extension\":\"xlsx\",\"order_by\":3,\"title\":\"\",\"display\":\"\",\"copyAsset\":false,\"role\":\"supplement\",\"size\":251459,\"visible\":true,\"origin\":\"\",\"legend\":\"Supplementary Data 3\",\"description\":\"\",\"filename\":\"SupplementaryData3.xlsx\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-9367379/v1/d5103168318f89feccb88272.xlsx\"},{\"id\":108189591,\"identity\":\"4d390e53-3fed-4fba-8e0a-1b7dba511032\",\"added_by\":\"auto\",\"created_at\":\"2026-04-30 09:42:18\",\"extension\":\"xlsx\",\"order_by\":4,\"title\":\"\",\"display\":\"\",\"copyAsset\":false,\"role\":\"supplement\",\"size\":20151,\"visible\":true,\"origin\":\"\",\"legend\":\"Supplementary Data 4\",\"description\":\"\",\"filename\":\"SupplementaryData4.xlsx\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-9367379/v1/15ce37c24456243909ddccca.xlsx\"},{\"id\":108189589,\"identity\":\"898adae3-8d41-4ac9-b853-b0948728ff4b\",\"added_by\":\"auto\",\"created_at\":\"2026-04-30 09:42:18\",\"extension\":\"xlsx\",\"order_by\":5,\"title\":\"\",\"display\":\"\",\"copyAsset\":false,\"role\":\"supplement\",\"size\":1196557,\"visible\":true,\"origin\":\"\",\"legend\":\"Supplementary Data 8\",\"description\":\"\",\"filename\":\"SupplementaryData8.xlsx\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-9367379/v1/3d7b91a6e31c2de50c91fb19.xlsx\"},{\"id\":108189607,\"identity\":\"059a38a2-fd58-463a-a45c-a5b16fa98288\",\"added_by\":\"auto\",\"created_at\":\"2026-04-30 09:42:22\",\"extension\":\"xlsx\",\"order_by\":6,\"title\":\"\",\"display\":\"\",\"copyAsset\":false,\"role\":\"supplement\",\"size\":18471,\"visible\":true,\"origin\":\"\",\"legend\":\"Supplementary Data 1\",\"description\":\"\",\"filename\":\"SupplementaryData1.xlsx\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-9367379/v1/bb7ed08f792e3d980b31aa87.xlsx\"},{\"id\":108189603,\"identity\":\"16bdca74-4379-44aa-96a8-07b931769daa\",\"added_by\":\"auto\",\"created_at\":\"2026-04-30 09:42:22\",\"extension\":\"xlsx\",\"order_by\":7,\"title\":\"\",\"display\":\"\",\"copyAsset\":false,\"role\":\"supplement\",\"size\":611822,\"visible\":true,\"origin\":\"\",\"legend\":\"Supplementary Data 5\",\"description\":\"\",\"filename\":\"SupplementaryData5.xlsx\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-9367379/v1/db64083bf1e778cee44077ed.xlsx\"},{\"id\":108189592,\"identity\":\"5d6cd7a9-595e-48ed-aab8-f047bca72b54\",\"added_by\":\"auto\",\"created_at\":\"2026-04-30 09:42:18\",\"extension\":\"xlsx\",\"order_by\":8,\"title\":\"\",\"display\":\"\",\"copyAsset\":false,\"role\":\"supplement\",\"size\":10848527,\"visible\":true,\"origin\":\"\",\"legend\":\"Supplementary Data 6\",\"description\":\"\",\"filename\":\"SupplementaryData6.xlsx\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-9367379/v1/fa5f63c32254e694e345363b.xlsx\"},{\"id\":108189608,\"identity\":\"3688d291-2254-4509-90a0-7238e0ddde25\",\"added_by\":\"auto\",\"created_at\":\"2026-04-30 09:42:22\",\"extension\":\"xlsx\",\"order_by\":9,\"title\":\"\",\"display\":\"\",\"copyAsset\":false,\"role\":\"supplement\",\"size\":11777337,\"visible\":true,\"origin\":\"\",\"legend\":\"Supplementary Data 7\",\"description\":\"\",\"filename\":\"SupplementaryData7.xlsx\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-9367379/v1/43ff6cba57a70ef736f79f23.xlsx\"}],\"financialInterests\":\"There is \\u003cb\\u003eNO\\u003c/b\\u003e Competing Interest.\",\"formattedTitle\":\"Mutual complementarity maximization in multi-view proteomics enables robust biomarker prediction\",\"fulltext\":[],\"fulltextSource\":\"\",\"fullText\":\"\",\"funders\":[],\"hasAdminPriorityOnWorkflow\":false,\"hasManuscriptDocX\":false,\"hasOptedInToPreprint\":true,\"hasPassedJournalQc\":\"\",\"hasAnyPriority\":true,\"hideJournal\":false,\"highlight\":\"\",\"institution\":\"\",\"isAcceptedByJournal\":false,\"isAuthorSuppliedPdf\":true,\"isDeskRejected\":\"\",\"isHiddenFromSearch\":false,\"isInQc\":false,\"isInWorkflow\":false,\"isPdf\":true,\"isPdfUpToDate\":true,\"isWithdrawnOrRetracted\":false,\"journal\":{\"display\":true,\"email\":\"info@researchsquare.com\",\"identity\":\"nature-portfolio\",\"isNatureJournal\":true,\"hasQc\":false,\"allowDirectSubmit\":false,\"externalIdentity\":\"\",\"sideBox\":\"\",\"snPcode\":\"\",\"submissionUrl\":\"\",\"title\":\"Nature Portfolio\",\"twitterHandle\":\"\",\"acdcEnabled\":false,\"dfaEnabled\":false,\"editorialSystem\":\"ejp\",\"reportingPortfolio\":\"\",\"inReviewEnabled\":true,\"inReviewRevisionsEnabled\":false},\"keywords\":\"Semi-supervised deep multi-view contrastive learning, Proteomics, Differentially expressed proteins, computational biology\",\"lastPublishedDoi\":\"10.21203/rs.3.rs-9367379/v1\",\"lastPublishedDoiUrl\":\"https://doi.org/10.21203/rs.3.rs-9367379/v1\",\"license\":{\"name\":\"CC BY 4.0\",\"url\":\"https://creativecommons.org/licenses/by/4.0/\"},\"manuscriptAbstract\":\"Mass spectrometry (MS)-based proteomic quantification is a multi-dimensional problem. Current approaches focus only on single dimensions (views), which is a limited representation of the underlying proteome. Distinct quantification modes (“views”) encodes complementary biological information, but each view is separately explored and not well-integrated. This has led towards biased single-view studies in comparative proteomic analysis, including the identification of differentially expressed proteins (DEPs), which are important for drug target discovery and disease diagnosis. To exploit mutual multi-view complementarity, we introduce Multi-view Contrastive Proteomics (MCP), a semi-supervised representation learning framework that integrates these heterogeneous quantification views into a shared embedding space through contrastive learning. MCP’s novelty lies in the explicit modeling of cross-view interactions while mitigating label scarcity, technical noise, and class imbalance. By jointly leveraging complementary views rather than treating them independently, MCP maximizes biologically meaningful data for DEP detection. Benchmarked on 18 gold-standard spike-in datasets, MCP consistently predicted more true DEPs, achieving \\u003e20% increase in recall while maintaining precision (\\u003e90%) in comparison with prevailing statistical and machine learning baselines. MCP also provides interpretable insights into cross-view interactions, revealing deeper structural mechanisms underpinning proteomic data. When applied to clinical proteomics cohorts, MCP unlocks previously inaccessible biological signals: robustly identifying unique DEPs that enhance thyroid nodule stratification. MCP also identified VSIG4, a macrophage immune suppressive regulator, as a novel prognostic biomarker for colorectal cancer that was missed by conventional approaches; in perturbation studies, MCP recovers more validated IFN-γ-responsive proteins and pathways from both bulk and single-cell proteomics, confirming the roles by perturbation modeling and virtual cell generation. Overall, MCP establishes a general computational paradigm for complementarity maximization in multi-view proteomic data analysis, enabling deeper biological understanding and translational impact. MCP is available at http://www.ai4pro.tech:3838.\",\"manuscriptTitle\":\"Mutual complementarity maximization in multi-view proteomics enables robust biomarker prediction\",\"msid\":\"\",\"msnumber\":\"\",\"nonDraftVersions\":[{\"code\":1,\"date\":\"2026-04-30 09:40:03\",\"doi\":\"10.21203/rs.3.rs-9367379/v1\",\"editorialEvents\":[],\"status\":\"published\",\"journal\":{\"display\":true,\"email\":\"info@researchsquare.com\",\"identity\":\"nature-communications\",\"isNatureJournal\":true,\"hasQc\":false,\"allowDirectSubmit\":false,\"externalIdentity\":\"NCOMMS\",\"sideBox\":\"Learn more about [Nature Communications](http://www.nature.com/ncomms/)\",\"snPcode\":\"\",\"submissionUrl\":\"https://mts-ncomms.nature.com/\",\"title\":\"Nature Communications\",\"twitterHandle\":\"\",\"acdcEnabled\":true,\"dfaEnabled\":true,\"editorialSystem\":\"ejp\",\"reportingPortfolio\":\"Nature Communications\",\"inReviewEnabled\":true,\"inReviewRevisionsEnabled\":false}}],\"origin\":\"\",\"ownerIdentity\":\"b07ea0e2-4e16-4b86-a939-8d1ad97b124a\",\"owner\":[],\"postedDate\":\"April 30th, 2026\",\"published\":true,\"recentEditorialEvents\":[],\"rejectedJournal\":[],\"revision\":\"\",\"amendment\":\"\",\"status\":\"under-review\",\"subjectAreas\":[{\"id\":66414556,\"name\":\"Biological sciences/Computational biology and bioinformatics/Computational models\"},{\"id\":66414557,\"name\":\"Biological sciences/Computational biology and bioinformatics/Proteome informatics\"}],\"tags\":[],\"updatedAt\":\"2026-04-30T09:40:03+00:00\",\"versionOfRecord\":[],\"versionCreatedAt\":\"2026-04-30 09:40:03\",\"video\":\"\",\"vorDoi\":\"\",\"vorDoiUrl\":\"\",\"workflowStages\":[]},\"version\":\"v1\",\"identity\":\"rs-9367379\",\"journalConfig\":\"researchsquare\"},\"__N_SSP\":true},\"page\":\"/article/[identity]/[[...version]]\",\"query\":{\"redirect\":\"/article/rs-9367379\",\"identity\":\"rs-9367379\",\"version\":[\"v1\"]},\"buildId\":\"XKTyCvWXoU3ODBz1xrDgd\",\"isFallback\":false,\"isExperimentalCompile\":false,\"dynamicIds\":[84888],\"gssp\":true,\"scriptLoader\":[]}","source_license":"CC-BY-4.0","license_restricted":false}