Multimodal deep learning integration of cryo-EM and AlphaFold3 for high-accuracy protein structure determination

Multimodal deep learning integration of cryo-EM and AlphaFold3 for high-accuracy protein structure determination | 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 Multimodal deep learning integration of cryo-EM and AlphaFold3 for high-accuracy protein structure determination Jianlin Cheng, Rajan Gyawali, Ashwin Dhakal This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7055373/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 31 Oct, 2025 Read the published version in Communications Chemistry → Version 1 posted You are reading this latest preprint version Abstract Cryo-electron microscopy (cryo-EM) is a key technology for determining the structures of proteins, particularly large protein complexes. However, automatically building high-accuracy protein structures from cryo-EM density maps remains a crucial challenge. In this work, we introduce MICA, a fully automatic and multimodal deep learning approach combining cryo-EM density maps with AlphaFold3-predicted structures at both input and output levels to improve cryo-EM protein structure modeling. It first uses a multi-task encoder-decoder architecture with a feature pyramid network to predict backbone atoms, Cα atoms and amino acid types from both cryo-EM maps and AlphaFold3-predicted structures, which are used to build an initial backbone model. This model is further refined using AlphaFold3-predicted structures and density maps to build final atomic structures. MICA significantly outperforms other state-of-the-art deep learning methods in terms of both modeling accuracy and completeness and is robust to protein size and map resolution. Additionally, it builds high-accuracy structural models with an average template-based modeling score (TM-score) of 0.93 from recently released high-resolution cryo-EM density maps, showing it can be used for real-world, automated, accurate protein structure determination. Biological sciences/Computational biology and bioinformatics/Machine learning Biological sciences/Structural biology/Electron microscopy/Cryoelectron microscopy Full Text Additional Declarations There is NO Competing Interest. Supplementary Files SupplementaryTables.pdf Supplementary document Cite Share Download PDF Status: Published Journal Publication published 31 Oct, 2025 Read the published version in Communications Chemistry → 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-7055373","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":482427828,"identity":"233288ce-ea9c-4f95-94c9-84e1547082ba","order_by":0,"name":"Jianlin Cheng","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA3UlEQVRIiWNgGAWjYDACCRDBIyHHwAzh8wCxAVFajEnVwsCQ2IAkhl8L/+zmYw+/yFikr21nf/zhY9s9GQb25m0SeC25cyzdWIZHInfbYR4zyZltxTwMPMfK8GoxkMgxk5aAaGFj5m1L4GEAihDQkv8NpCXd7DD7489gLfJvCGnJYZP8wCORYHaYwUAaYgsPfi0SN9LMpIGBbAj2y4xzCTxsPGnFFvi08M9Ifib5s6dO3uz88ccfPpQl2POzH954A58WEGDm7UHisRFSDgKMP34Qo2wUjIJRMApGLAAAWOY8ybNhuQUAAAAASUVORK5CYII=","orcid":"https://orcid.org/0000-0003-0305-2853","institution":"University of Missouri","correspondingAuthor":true,"prefix":"","firstName":"Jianlin","middleName":"","lastName":"Cheng","suffix":""},{"id":482427829,"identity":"ba1c1a90-e6be-49da-8048-1ca0b70794da","order_by":1,"name":"Rajan Gyawali","email":"","orcid":"https://orcid.org/0000-0002-7052-4964","institution":"University of Missouri-Columbia","correspondingAuthor":false,"prefix":"","firstName":"Rajan","middleName":"","lastName":"Gyawali","suffix":""},{"id":482427830,"identity":"50a6453e-c262-4689-a0e2-5a112ef9b0f8","order_by":2,"name":"Ashwin Dhakal","email":"","orcid":"","institution":"University of Missouri - Columbia","correspondingAuthor":false,"prefix":"","firstName":"Ashwin","middleName":"","lastName":"Dhakal","suffix":""}],"badges":[],"createdAt":"2025-07-06 02:05:14","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7055373/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7055373/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1038/s42004-025-01718-5","type":"published","date":"2025-10-31T04:00:00+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":94903774,"identity":"0acfbd01-fa99-4642-8887-691f9efeb4ea","added_by":"auto","created_at":"2025-11-01 07:07:14","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2236655,"visible":true,"origin":"","legend":"","description":"","filename":"MICAmanuscriptv1.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7055373/v1_covered_bfcef920-6817-42d3-b27e-29a9339f0426.pdf"},{"id":87331318,"identity":"0ac3c24c-0398-4981-aec5-63d555d86f42","added_by":"auto","created_at":"2025-07-22 19:07:51","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":1108702,"visible":true,"origin":"","legend":"Supplementary document","description":"","filename":"SupplementaryTables.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7055373/v1/a557bfefcd89c894ad28e93c.pdf"}],"financialInterests":"There is \u003cb\u003eNO\u003c/b\u003e Competing Interest.","formattedTitle":"Multimodal deep learning integration of cryo-EM and AlphaFold3 for high-accuracy protein structure determination","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-7055373/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7055373/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eCryo-electron microscopy (cryo-EM) is a key technology for determining the structures of proteins, particularly large protein complexes. However, automatically building high-accuracy protein structures from cryo-EM density maps remains a crucial challenge. In this work, we introduce MICA, a fully automatic and multimodal deep learning approach combining cryo-EM density maps with AlphaFold3-predicted structures at both input and output levels to improve cryo-EM protein structure modeling. It first uses a multi-task encoder-decoder architecture with a feature pyramid network to predict backbone atoms, Cα atoms and amino acid types from both cryo-EM maps and AlphaFold3-predicted structures, which are used to build an initial backbone model. This model is further refined using AlphaFold3-predicted structures and density maps to build final atomic structures. MICA significantly outperforms other state-of-the-art deep learning methods in terms of both modeling accuracy and completeness and is robust to protein size and map resolution. Additionally, it builds high-accuracy structural models with an average template-based modeling score (TM-score) of 0.93 from recently released high-resolution cryo-EM density maps, showing it can be used for real-world, automated, accurate protein structure determination.\u003c/p\u003e","manuscriptTitle":"Multimodal deep learning integration of cryo-EM and AlphaFold3 for high-accuracy protein structure determination","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-07-22 18:59:46","doi":"10.21203/rs.3.rs-7055373/v1","editorialEvents":[],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"communications-chemistry","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"commschem","sideBox":"Learn more about [Communications Chemistry](http://www.nature.com/commschem/)","snPcode":"","submissionUrl":"","title":"Communications Chemistry","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"Communications Series","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"f097f0a9-f915-426a-ba65-a7f3ffad58a5","owner":[],"postedDate":"July 22nd, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[{"id":51223080,"name":"Biological sciences/Computational biology and bioinformatics/Machine learning"},{"id":51223081,"name":"Biological sciences/Structural biology/Electron microscopy/Cryoelectron microscopy"}],"tags":[],"updatedAt":"2025-11-01T07:07:07+00:00","versionOfRecord":{"articleIdentity":"rs-7055373","link":"https://doi.org/10.1038/s42004-025-01718-5","journal":{"identity":"communications-chemistry","isVorOnly":false,"title":"Communications Chemistry"},"publishedOn":"2025-10-31 04:00:00","publishedOnDateReadable":"October 31st, 2025"},"versionCreatedAt":"2025-07-22 18:59:46","video":"","vorDoi":"10.1038/s42004-025-01718-5","vorDoiUrl":"https://doi.org/10.1038/s42004-025-01718-5","workflowStages":[]},"version":"v1","identity":"rs-7055373","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7055373","identity":"rs-7055373","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}