Accelerating protein directed evolution via reinforcement learning

Accelerating protein directed evolution via reinforcement 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 Accelerating protein directed evolution via reinforcement learning Haipeng Gong, Tianyu Mi, Yu-Xiang Wang, Wanze Wang, Jingyu Zhao, and 6 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8907793/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 With the advancement of artificial intelligence, the protein fitness landscape becomes predictable, providing reliable guidance in the selection of advantageous mutations for the directed evolution of proteins. In practice, however, simply combining a small number of advantageous single mutations is unlikely to produce variants with global superiority, while exhaustive exploration of the astronomical mutational combinations is highly challenging. In this study, we introduce a virtual directed evolution pipeline, RelaVDEP, to facilitate functional optimization of the target protein in silico. By developing a reward model to balance the efficiency and accuracy of protein functional prediction and by designing a model-based reinforcement learning framework to explore the vast combinatorial space of protein mutations, this pipeline is capable of automatically identifying diversified multiple mutational variants with notable improvement in desired functional properties for a broad spectrum of proteins (including highly engineered targets like eGFP and PETase), as evidenced by experimental validation Biological sciences/Computational biology and bioinformatics/Computational models Biological sciences/Computational biology and bioinformatics/Machine learning protein engineering directed evolution mutational effects reinforcement learning sequence optimization Full Text Additional Declarations There is NO Competing Interest. 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-8907793","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":599484517,"identity":"45de7277-9910-43b9-8ee4-25a9998f0e96","order_by":0,"name":"Haipeng Gong","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA7klEQVRIiWNgGAWjYBACAzBZwcDABkYgcIAoLWcMSNXC2AamiNRizr/42MOv8/7Y80m3X3vMU8Mgx3cjgfFzAR4tljOepRvLbjNIbJM5U27Mc4zBWPJGArP0DHwOu3HGTFpym0ECm0ROmjQPG0PihhsJbMw8BLXMMbCHaPnHUE9Yy/keM8mPDQaMbRLpx6R52xgSDAjbwpYmzXDMOLFNIodNcm6fhOHMMw+bpfHbcviY5I8aOXv5GenPJN58s5HnO5588DM+LQwSCQxQZ/CAIkcCiBkb8GlgYOA/wMD4A8xif4Bf5SgYBaNgFIxYAADmTEf5u1iCcAAAAABJRU5ErkJggg==","orcid":"https://orcid.org/0000-0002-5532-1640","institution":"Tsinghua University","correspondingAuthor":true,"prefix":"","firstName":"Haipeng","middleName":"","lastName":"Gong","suffix":""},{"id":599484518,"identity":"01db57dd-c354-4c97-9ac4-e29c396142b1","order_by":1,"name":"Tianyu Mi","email":"","orcid":"","institution":"Tsinghua University","correspondingAuthor":false,"prefix":"","firstName":"Tianyu","middleName":"","lastName":"Mi","suffix":""},{"id":599484519,"identity":"75a65074-767a-4b29-b3e4-da4b1aae247b","order_by":2,"name":"Yu-Xiang Wang","email":"","orcid":"","institution":"Peking Univeristy","correspondingAuthor":false,"prefix":"","firstName":"Yu-Xiang","middleName":"","lastName":"Wang","suffix":""},{"id":599484520,"identity":"d4d42fdf-c6f7-4530-9692-059e6804c02a","order_by":3,"name":"Wanze Wang","email":"","orcid":"","institution":"Tsinghua University","correspondingAuthor":false,"prefix":"","firstName":"Wanze","middleName":"","lastName":"Wang","suffix":""},{"id":599484521,"identity":"943bc332-77c3-4499-a3f4-818588864923","order_by":4,"name":"Jingyu Zhao","email":"","orcid":"","institution":"Tsinghua University","correspondingAuthor":false,"prefix":"","firstName":"Jingyu","middleName":"","lastName":"Zhao","suffix":""},{"id":599484522,"identity":"dfb8474b-fe0c-4ebb-9562-529eff423115","order_by":5,"name":"Yunhao Shen","email":"","orcid":"","institution":"Tsinghua University","correspondingAuthor":false,"prefix":"","firstName":"Yunhao","middleName":"","lastName":"Shen","suffix":""},{"id":599484523,"identity":"4161892d-54a8-47cb-b76a-a269a8eaf4d5","order_by":6,"name":"Nan Xiao","email":"","orcid":"","institution":"Tsinghua University","correspondingAuthor":false,"prefix":"","firstName":"Nan","middleName":"","lastName":"Xiao","suffix":""},{"id":599484524,"identity":"9356903f-0f66-4fac-82f4-2620821b890f","order_by":7,"name":"Ligong Chen","email":"","orcid":"https://orcid.org/0000-0002-7893-7173","institution":"Tsinghua University","correspondingAuthor":false,"prefix":"","firstName":"Ligong","middleName":"","lastName":"Chen","suffix":""},{"id":599484525,"identity":"bf660c2a-61ca-44f3-a0c5-872598aa7d35","order_by":8,"name":"Guo-Qiang Chen","email":"","orcid":"https://orcid.org/0000-0002-7226-1782","institution":"Tsinghua university","correspondingAuthor":false,"prefix":"","firstName":"Guo-Qiang","middleName":"","lastName":"Chen","suffix":""},{"id":599484526,"identity":"96c35bd2-3150-4f3b-a357-2e40cdb66b6b","order_by":9,"name":"Shuyi Zhang","email":"","orcid":"https://orcid.org/0000-0001-8500-5836","institution":"Tsinghua University","correspondingAuthor":false,"prefix":"","firstName":"Shuyi","middleName":"","lastName":"Zhang","suffix":""},{"id":599484527,"identity":"dd65f48e-1399-4376-97d5-26c61091003b","order_by":10,"name":"Wen-Bin Zhang","email":"","orcid":"https://orcid.org/0000-0002-8746-0792","institution":"Peking University","correspondingAuthor":false,"prefix":"","firstName":"Wen-Bin","middleName":"","lastName":"Zhang","suffix":""}],"badges":[],"createdAt":"2026-02-18 09:27:47","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8907793/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8907793/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":107483693,"identity":"7d95cad0-b0ee-49ef-87a6-c8f9b30500d6","added_by":"auto","created_at":"2026-04-22 02:28:51","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":19737522,"visible":true,"origin":"","legend":"Article File","description":"","filename":"RelaVDEP.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8907793/v1_covered_dca06096-fc84-48bf-a73a-7fde3ba3f441.pdf"}],"financialInterests":"There is \u003cb\u003eNO\u003c/b\u003e Competing Interest.","formattedTitle":"Accelerating protein directed evolution via reinforcement learning","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":"[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":"protein engineering, directed evolution, mutational effects, reinforcement learning, sequence optimization","lastPublishedDoi":"10.21203/rs.3.rs-8907793/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8907793/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"With the advancement of artificial intelligence, the protein fitness landscape becomes predictable, providing reliable guidance in the selection of advantageous mutations for the directed evolution of proteins. In practice, however, simply combining a small number of advantageous single mutations is unlikely to produce variants with global superiority, while exhaustive exploration of the astronomical mutational combinations is highly challenging. In this study, we introduce a virtual directed evolution pipeline, RelaVDEP, to facilitate functional optimization of the target protein in silico. By developing a reward model to balance the efficiency and accuracy of protein functional prediction and by designing a model-based reinforcement learning framework to explore the vast combinatorial space of protein mutations, this pipeline is capable of automatically identifying diversified multiple mutational variants with notable improvement in desired functional properties for a broad spectrum of proteins (including highly engineered targets like eGFP and PETase), as evidenced by experimental validation","manuscriptTitle":"Accelerating protein directed evolution via reinforcement learning","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-03-04 05:51:33","doi":"10.21203/rs.3.rs-8907793/v1","editorialEvents":[],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"nature-methods","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"nmeth","sideBox":"Learn more about [Nature Methods](http://www.nature.com/nmeth)","snPcode":"","submissionUrl":"","title":"Nature Methods","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"Nature Research","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"48e78e90-b793-4966-af7c-60d933563fb2","owner":[],"postedDate":"March 4th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[{"id":63790464,"name":"Biological sciences/Computational biology and bioinformatics/Computational models"},{"id":63790465,"name":"Biological sciences/Computational biology and bioinformatics/Machine learning"}],"tags":[],"updatedAt":"2026-05-06T22:00:15+00:00","versionOfRecord":[],"versionCreatedAt":"2026-03-04 05:51:33","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8907793","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8907793","identity":"rs-8907793","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}