Improving generalizability and data efficiency for MHC-I binding peptide predictions through structure-based geometric deep learning

Improving generalizability and data efficiency for MHC-I binding peptide predictions through structure-based geometric deep learning | 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 Improving generalizability and data efficiency for MHC-I binding peptide predictions through structure-based geometric deep learning Li Xue, Dario Marzella, Giulia Crocioni, Tadija Radusinović, Daniil Lepikhov, and 10 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-3924124/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 19 Dec, 2024 Read the published version in Communications Biology → Version 1 posted You are reading this latest preprint version Abstract The interaction between peptides and major histocompatibility complex (MHC) molecules is pivotal for tissue transplantation, pathogen recognition and autoimmune disease treatments. Recent advances in cancer immunotherapies demand for more accurate computational prediction of MHC-bound peptides. We address the generalizability challenge of MHC-bound peptide predictions, revealing limitations in current sequence-based approaches. Our solution employs structure-based methods leveraging geometric deep learning (GDL), yielding up to 8% improvement in generalizability across unseen MHC alleles. We tackle data efficiency by introducing a self-supervised learning approach surpassing sequence-based methods, even without being trained on binding affinity data. Finally, we demonstrate the resilience of structure-based GDL methods to biases in binding data on an Hepatitis B virus vaccine design case study. This study highlights structure-based methods’ potential to enhance generalizability and data efficiency, with implications for data-intensive fields like T-cell receptor specificity predictions, paving the way for enhanced comprehension and manipulation of immune responses. Biological sciences/Immunology/Antigen processing and presentation Biological sciences/Biophysics/Computational biophysics Biological sciences/Computational biology and bioinformatics/Machine learning Biological sciences/Computational biology and bioinformatics/Computational models Full Text Additional Declarations There is NO Competing Interest. Supplementary Files 3dvacsupplmat.pdf SupplVideo.mov Supplementary Video 1 Cite Share Download PDF Status: Published Journal Publication published 19 Dec, 2024 Read the published version in Communications Biology → 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-3924124","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":279128162,"identity":"9af3f675-245c-45a8-bc42-cf7cbad5f793","order_by":0,"name":"Li Xue","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAxUlEQVRIiWNgGAWjYBACPmYwZQPEOUDMBuYxHsCnhQ2iJQ1FCwN+LRDqMCla2BnYpHkqzufzt+cefMxTdjiav4H3ACGHMRvznLltOePMu2RjnnOHc2cc4EsgpIXxMW/bbQOGGzlm0rxth3M3MPAYENLCcJj33zkDeVK0AG1pOGBgQIIWxmbDOceSDQyBfjGccy49d8ZhAlr4+Q8fk3hTY2cgdzz34IM3Zda5/e09hg/waQFGdQOaADN+9aNgFIyCUTAKiAAAmERDyJKmxZYAAAAASUVORK5CYII=","orcid":"","institution":"Radboud University Medical Center","correspondingAuthor":true,"prefix":"","firstName":"Li","middleName":"","lastName":"Xue","suffix":""},{"id":279128163,"identity":"2bc5f0a0-14af-4616-aa45-785d29f01934","order_by":1,"name":"Dario Marzella","email":"","orcid":"","institution":"Radboud University Medical Center","correspondingAuthor":false,"prefix":"","firstName":"Dario","middleName":"","lastName":"Marzella","suffix":""},{"id":279128164,"identity":"45ec54b8-c298-465d-b227-12fb368e9466","order_by":2,"name":"Giulia Crocioni","email":"","orcid":"","institution":"Netherlands eScience Center","correspondingAuthor":false,"prefix":"","firstName":"Giulia","middleName":"","lastName":"Crocioni","suffix":""},{"id":279128165,"identity":"e2ceea22-53f3-44d2-9437-67984559bb41","order_by":3,"name":"Tadija Radusinović","email":"","orcid":"","institution":"University of Amsterdam","correspondingAuthor":false,"prefix":"","firstName":"Tadija","middleName":"","lastName":"Radusinović","suffix":""},{"id":279128166,"identity":"f290ffb7-5638-4057-9a15-9427e91d7ab3","order_by":4,"name":"Daniil Lepikhov","email":"","orcid":"https://orcid.org/0009-0001-1797-0850","institution":"Radboud University Medical Center","correspondingAuthor":false,"prefix":"","firstName":"Daniil","middleName":"","lastName":"Lepikhov","suffix":""},{"id":279128167,"identity":"30aa0eaf-f036-4ef0-8627-22c585518928","order_by":5,"name":"Heleen Severin","email":"","orcid":"","institution":"Radboud University Medical Center","correspondingAuthor":false,"prefix":"","firstName":"Heleen","middleName":"","lastName":"Severin","suffix":""},{"id":279128168,"identity":"ee4bb132-6546-437e-b139-5352886be229","order_by":6,"name":"Dani Bodor","email":"","orcid":"","institution":"Netherlans eScience Center","correspondingAuthor":false,"prefix":"","firstName":"Dani","middleName":"","lastName":"Bodor","suffix":""},{"id":279128169,"identity":"ac4f1ca6-56b0-448e-8046-a485ff2a7e80","order_by":7,"name":"Daniel Rademaker","email":"","orcid":"","institution":"Radboud University Medical Center","correspondingAuthor":false,"prefix":"","firstName":"Daniel","middleName":"","lastName":"Rademaker","suffix":""},{"id":279128170,"identity":"f212ddca-8c9a-47d9-b82f-868d8a1f92cf","order_by":8,"name":"ChiaYu Lin","email":"","orcid":"","institution":"Netherlands eScience Center","correspondingAuthor":false,"prefix":"","firstName":"ChiaYu","middleName":"","lastName":"Lin","suffix":""},{"id":279128171,"identity":"b6f66040-abe6-4eb5-b441-5772cdd1c6b0","order_by":9,"name":"Sonja Georgievska","email":"","orcid":"https://orcid.org/0000-0002-8094-4532","institution":"Netherlands eScience Center","correspondingAuthor":false,"prefix":"","firstName":"Sonja","middleName":"","lastName":"Georgievska","suffix":""},{"id":279128172,"identity":"2af044ea-6414-4db5-8ea6-856256f6098b","order_by":10,"name":"Nicolas Renaud","email":"","orcid":"","institution":"Netherlands eScience Center","correspondingAuthor":false,"prefix":"","firstName":"Nicolas","middleName":"","lastName":"Renaud","suffix":""},{"id":279128173,"identity":"6c0c0038-23a7-4c1c-95bd-c173c78a766e","order_by":11,"name":"Amy Kessler","email":"","orcid":"","institution":"Erasmus MC, University Medical Center Rotterdam","correspondingAuthor":false,"prefix":"","firstName":"Amy","middleName":"","lastName":"Kessler","suffix":""},{"id":279128174,"identity":"cac4ef60-7c70-4a68-b913-07955c9e19f6","order_by":12,"name":"Pablo Lopez-Tarifa","email":"","orcid":"","institution":"Netherlands eScience Center","correspondingAuthor":false,"prefix":"","firstName":"Pablo","middleName":"","lastName":"Lopez-Tarifa","suffix":""},{"id":279128175,"identity":"75d4c910-875d-4a91-826b-571788653ef5","order_by":13,"name":"Sonja Buschow","email":"","orcid":"","institution":"Erasmus MC, University Medical Center Rotterdam","correspondingAuthor":false,"prefix":"","firstName":"Sonja","middleName":"","lastName":"Buschow","suffix":""},{"id":279128176,"identity":"abc89040-aa0c-4885-a066-ea3b2a72ccf0","order_by":14,"name":"Erik Bekkers","email":"","orcid":"","institution":"University of Amsterdam","correspondingAuthor":false,"prefix":"","firstName":"Erik","middleName":"","lastName":"Bekkers","suffix":""}],"badges":[],"createdAt":"2024-02-03 13:45:10","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-3924124/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-3924124/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1038/s42003-024-07292-1","type":"published","date":"2024-12-19T05:00:00+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":71970290,"identity":"e64364a9-bd44-43a8-ad71-a06902ead07d","added_by":"auto","created_at":"2024-12-20 08:09:25","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1285576,"visible":true,"origin":"","legend":"","description":"","filename":"3dvacEGNNSSL.pdf","url":"https://assets-eu.researchsquare.com/files/rs-3924124/v1_covered_ca8f700b-e1f7-44f6-ae81-d821698ebcf0.pdf"},{"id":53443886,"identity":"1aa38df3-5530-49f8-8c5a-67a43e81fefc","added_by":"auto","created_at":"2024-03-26 04:23:57","extension":"pdf","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":986942,"visible":true,"origin":"","legend":"","description":"","filename":"3dvacsupplmat.pdf","url":"https://assets-eu.researchsquare.com/files/rs-3924124/v1/913826ac108fb92ae3f610df.pdf"},{"id":53443887,"identity":"2b4b2849-7a32-42d9-89b0-a3428f717868","added_by":"auto","created_at":"2024-03-26 04:23:57","extension":"mov","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":5770858,"visible":true,"origin":"","legend":"Supplementary Video 1","description":"","filename":"SupplVideo.mov","url":"https://assets-eu.researchsquare.com/files/rs-3924124/v1/ba3411a1ca15d0e295613f97.mov"}],"financialInterests":"There is \u003cb\u003eNO\u003c/b\u003e Competing Interest.","formattedTitle":"Improving generalizability and data efficiency for MHC-I binding peptide predictions through structure-based geometric deep learning","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":true,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":true,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","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":"","lastPublishedDoi":"10.21203/rs.3.rs-3924124/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-3924124/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"The interaction between peptides and major histocompatibility complex (MHC) molecules is pivotal for tissue transplantation, pathogen recognition and autoimmune disease treatments. Recent advances in cancer immunotherapies demand for more accurate computational prediction of MHC-bound peptides. We address the generalizability challenge of MHC-bound peptide predictions, revealing limitations in current sequence-based approaches. Our solution employs structure-based methods leveraging geometric deep learning (GDL), yielding up to 8% improvement in generalizability across unseen MHC alleles. We tackle data efficiency by introducing a self-supervised learning approach surpassing sequence-based methods, even without being trained on binding affinity data. Finally, we demonstrate the resilience of structure-based GDL methods to biases in binding data on an Hepatitis B virus vaccine design case study. This study highlights structure-based methods’ potential to enhance generalizability and data efficiency, with implications for data-intensive fields like T-cell receptor specificity predictions, paving the way for enhanced comprehension and manipulation of immune responses.","manuscriptTitle":"Improving generalizability and data efficiency for MHC-I binding peptide predictions through structure-based geometric deep learning","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-03-26 04:23:52","doi":"10.21203/rs.3.rs-3924124/v1","editorialEvents":[],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"communications-biology","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"commsbio","sideBox":"Learn more about [Communications Biology](http://www.nature.com/commsbio/)","snPcode":"","submissionUrl":"","title":"Communications Biology","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"Communications Series","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"84221dbe-e56f-45f7-812a-8d1ff3a7246c","owner":[],"postedDate":"March 26th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[{"id":29396032,"name":"Biological sciences/Immunology/Antigen processing and presentation"},{"id":29396033,"name":"Biological sciences/Biophysics/Computational biophysics"},{"id":29396034,"name":"Biological sciences/Computational biology and bioinformatics/Machine learning"},{"id":29396035,"name":"Biological sciences/Computational biology and bioinformatics/Computational models"}],"tags":[],"updatedAt":"2024-12-20T08:09:14+00:00","versionOfRecord":{"articleIdentity":"rs-3924124","link":"https://doi.org/10.1038/s42003-024-07292-1","journal":{"identity":"communications-biology","isVorOnly":false,"title":"Communications Biology"},"publishedOn":"2024-12-19 05:00:00","publishedOnDateReadable":"December 19th, 2024"},"versionCreatedAt":"2024-03-26 04:23:52","video":"","vorDoi":"10.1038/s42003-024-07292-1","vorDoiUrl":"https://doi.org/10.1038/s42003-024-07292-1","workflowStages":[]},"version":"v1","identity":"rs-3924124","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-3924124","identity":"rs-3924124","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}