An SE(3)-equivariant language model for pocket-aware 3D molecular generation enables discovery of potent HPK1 inhibitors

An SE(3)-equivariant language model for pocket-aware 3D molecular generation enables discovery of potent HPK1 inhibitors | 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 An SE(3)-equivariant language model for pocket-aware 3D molecular generation enables discovery of potent HPK1 inhibitors Bo Huang, Bin Xi, Han Wang, Zhenming Liu, Guanglong Sun, Huting Wang, and 11 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8880270/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Generating molecules that simultaneously achieve optimal 3D pocket-binding conformations and chemically plausible topologies remains a central challenge in AI for Structure-Based Drug Design. Graph-based models excel in SE(3)-equivariant spatial reasoning but often struggle to ensure chemical validity, whereas language models capture discrete chemical syntax yet lack 3D spatial understanding. Here we introduce SE3-BiLingoMol, an SE(3)-equivariant language model for pocket-aware 3D ligand de novo generation and fragment-guided optimization. Built upon Geometric Algebra Transformers and a fragment-aware SMILES representation, our model enables SE(3)-equivariant modeling of continuous 3D geometry while ensuring chemically valid molecular topologies. To counteract cumulative 3D conformational errors inherent to autoregressive generation, we developed a bidirectional attention-based self-refinement mechanism as a key architectural component of SE3-BiLingoMol. Our model achieves state-of-the-art performance in an in-silico evaluation across over 100 diverse targets. Critically, application of SE3-BiLingoMol led to the discovery of a novel tetracyclic HPK1 inhibitor showing potent in vitro activity and robust in vivo anti-tumor efficacy. This work demonstrates a powerful and practical generative AI framework for accelerating structure-based drug design. Biological sciences/Drug discovery/Medicinal chemistry/Structure-based drug design Biological sciences/Chemical biology/Small molecules Biological sciences/Computational biology and bioinformatics/Machine learning Health sciences/Diseases/Cancer/Cancer therapy/Cancer immunotherapy Biological sciences/Drug discovery/Drug screening/Virtual screening Full Text Additional Declarations There is NO Competing Interest. Supplementary Files NCSIv1.pdf Supplementary Information D1300062751valreportannotateP1.pdf X-ray Structure Validation Summary Report HPK1Cmpd2.txt Model file for evaluation Cite Share Download PDF Status: Posted 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-8880270","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":597272827,"identity":"ab0020b1-5ab2-4f9d-a28f-5ffe1a9b0ffd","order_by":0,"name":"Bo Huang","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAsUlEQVRIiWNgGAWjYBAC9gYGhgNAWo6Nvf0AcVp4gOoOALExH8+ZBOK1gKxJnCfhYECkFukzhoc/VBxOb5NgSGD4UbGNCC18OQYHDpw5nNsm3XiAsefMbcJa7Hl4DA4cbANqkTmQwMzYRoQWHqiWdDaJBAPStCSQooWt4MCZM+mGbcBAPkiUX3h4mDd/qKiwlpdvbz/44EcFEVqgoBlMHiBaPRDUkaJ4FIyCUTAKRhoAANwfPZdt9Ot1AAAAAElFTkSuQmCC","orcid":"https://orcid.org/0000-0003-3822-9110","institution":"Capital Medical University","correspondingAuthor":true,"prefix":"","firstName":"Bo","middleName":"","lastName":"Huang","suffix":""},{"id":597272828,"identity":"b0543b6b-817f-47c4-80e5-87ff7b735ee9","order_by":1,"name":"Bin Xi","email":"","orcid":"https://orcid.org/0000-0003-0141-2157","institution":"Peking University","correspondingAuthor":false,"prefix":"","firstName":"Bin","middleName":"","lastName":"Xi","suffix":""},{"id":597272829,"identity":"2ee6142a-3c7e-4d21-847a-9f23244f76b7","order_by":2,"name":"Han Wang","email":"","orcid":"https://orcid.org/0000-0001-9282-2105","institution":"State Key Laboratory of Natural and Biomimetic Drugs, School of Pharmaceutical Sciences, Peking University","correspondingAuthor":false,"prefix":"","firstName":"Han","middleName":"","lastName":"Wang","suffix":""},{"id":597272830,"identity":"604cb26a-df4e-4fc4-b5ea-fae76897540d","order_by":3,"name":"Zhenming Liu","email":"","orcid":"https://orcid.org/0000-0002-8993-4015","institution":"Peking University","correspondingAuthor":false,"prefix":"","firstName":"Zhenming","middleName":"","lastName":"Liu","suffix":""},{"id":597272831,"identity":"60e2aa88-043b-4a3e-9e4b-4943faf73ed9","order_by":4,"name":"Guanglong Sun","email":"","orcid":"","institution":"Beijing StoneWise Technology Co Ltd.","correspondingAuthor":false,"prefix":"","firstName":"Guanglong","middleName":"","lastName":"Sun","suffix":""},{"id":597272832,"identity":"da4b3b9f-5453-4e39-aed4-e7475957f7a8","order_by":5,"name":"Huting Wang","email":"","orcid":"","institution":"Beijing StoneWise Technology Co Ltd.","correspondingAuthor":false,"prefix":"","firstName":"Huting","middleName":"","lastName":"Wang","suffix":""},{"id":597272833,"identity":"fe5b7674-2df0-44e9-8b55-287e5bdc7cb5","order_by":6,"name":"Ruihan Mao","email":"","orcid":"","institution":"Beijing StoneWise Technology Co Ltd.","correspondingAuthor":false,"prefix":"","firstName":"Ruihan","middleName":"","lastName":"Mao","suffix":""},{"id":597272834,"identity":"7a60d62d-5af0-4a63-a855-f76e59366939","order_by":7,"name":"Yuyang Ge","email":"","orcid":"","institution":"Beijing StoneWise Technology Co Ltd.","correspondingAuthor":false,"prefix":"","firstName":"Yuyang","middleName":"","lastName":"Ge","suffix":""},{"id":597272835,"identity":"741d1cc8-f9ff-4039-a9e6-c1077f3f2c8e","order_by":8,"name":"Yang Wang","email":"","orcid":"","institution":"Beijing StoneWise Technology Co Ltd.","correspondingAuthor":false,"prefix":"","firstName":"Yang","middleName":"","lastName":"Wang","suffix":""},{"id":597272836,"identity":"b4f938e5-8f8a-45d9-81c5-46938ee536a6","order_by":9,"name":"Feng Zhou","email":"","orcid":"https://orcid.org/0000-0001-9040-1796","institution":"Beijing StoneWise Technology Co Ltd","correspondingAuthor":false,"prefix":"","firstName":"Feng","middleName":"","lastName":"Zhou","suffix":""},{"id":597272837,"identity":"5451a23c-358c-4fe2-b164-58cc62c385a6","order_by":10,"name":"Wenbiao Zhou","email":"","orcid":"https://orcid.org/0000-0002-7168-3676","institution":"Beijing StoneWise Technology Co Ltd.","correspondingAuthor":false,"prefix":"","firstName":"Wenbiao","middleName":"","lastName":"Zhou","suffix":""},{"id":597272838,"identity":"bdde26ec-fef5-41e7-9e71-fcad8fa5b96e","order_by":11,"name":"Yiting Pan","email":"","orcid":"","institution":"Beijing StoneWise Technology Co Ltd.","correspondingAuthor":false,"prefix":"","firstName":"Yiting","middleName":"","lastName":"Pan","suffix":""},{"id":597272839,"identity":"afd0a19a-d009-4a91-b459-17d569745187","order_by":12,"name":"Bowen Zhang","email":"","orcid":"","institution":"Peking University","correspondingAuthor":false,"prefix":"","firstName":"Bowen","middleName":"","lastName":"Zhang","suffix":""},{"id":597272840,"identity":"b0d138c6-5a4f-4a65-8ddc-18fa11e6e591","order_by":13,"name":"Jiangtao Zhang","email":"","orcid":"","institution":"Laboratory of Soft Matter Physics, Institute of Physics, Chinese Academy of Sciences","correspondingAuthor":false,"prefix":"","firstName":"Jiangtao","middleName":"","lastName":"Zhang","suffix":""},{"id":597272841,"identity":"49aa2125-43ec-431f-9638-087e9cb8a369","order_by":14,"name":"Daohua Jiang","email":"","orcid":"https://orcid.org/0000-0002-4397-5807","institution":"Institute of Physics, Chinese Academy of Sciences","correspondingAuthor":false,"prefix":"","firstName":"Daohua","middleName":"","lastName":"Jiang","suffix":""},{"id":597272842,"identity":"ff1e6b17-76a2-44fa-b326-394710017622","order_by":15,"name":"Yuji Wang","email":"","orcid":"","institution":"Department of Medicinal Chemistry, College of Pharmaceutical Sciences of Capital Medical University","correspondingAuthor":false,"prefix":"","firstName":"Yuji","middleName":"","lastName":"Wang","suffix":""},{"id":597272843,"identity":"7085cb49-7d25-4ac1-bf63-a8066d00f3dc","order_by":16,"name":"Xiaojian Xu","email":"","orcid":"","institution":"Beijing Neurosurgical Institute","correspondingAuthor":false,"prefix":"","firstName":"Xiaojian","middleName":"","lastName":"Xu","suffix":""}],"badges":[],"createdAt":"2026-02-14 13:25:12","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8880270/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8880270/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":109068183,"identity":"6e69f009-6448-45c8-b3c2-e8e352cd4527","added_by":"auto","created_at":"2026-05-12 10:04:21","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":3066152,"visible":true,"origin":"","legend":"Article File","description":"","filename":"NCarticle.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8880270/v1_covered_c94dd5a6-9446-4d7a-a3c9-6ac50c181345.pdf"},{"id":103556134,"identity":"68f4ba75-622d-4ec5-94a7-e5379863effd","added_by":"auto","created_at":"2026-02-27 04:04:44","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":1677589,"visible":true,"origin":"","legend":"Supplementary Information","description":"","filename":"NCSIv1.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8880270/v1/9b488850c23079ae9070db35.pdf"},{"id":103556135,"identity":"1ab99e79-adee-4214-9d6a-43e388e123bc","added_by":"auto","created_at":"2026-02-27 04:04:44","extension":"pdf","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":3324732,"visible":true,"origin":"","legend":"X-ray Structure Validation Summary Report","description":"","filename":"D1300062751valreportannotateP1.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8880270/v1/f9face77fb0de526b3a07a14.pdf"},{"id":103556136,"identity":"6e87ce5f-84db-4a86-85dd-1e5fd9ca15d0","added_by":"auto","created_at":"2026-02-27 04:04:44","extension":"txt","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":2857872,"visible":true,"origin":"","legend":"Model file for evaluation","description":"","filename":"HPK1Cmpd2.txt","url":"https://assets-eu.researchsquare.com/files/rs-8880270/v1/4564ed576cfa919f14ce4cf1.txt"}],"financialInterests":"There is \u003cb\u003eNO\u003c/b\u003e Competing Interest.","formattedTitle":"An SE(3)-equivariant language model for pocket-aware 3D molecular generation enables discovery of potent HPK1 inhibitors","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"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":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-8880270/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8880270/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"Generating molecules that simultaneously achieve optimal 3D pocket-binding conformations and chemically plausible topologies remains a central challenge in AI for Structure-Based Drug Design. Graph-based models excel in SE(3)-equivariant spatial reasoning but often struggle to ensure chemical validity, whereas language models capture discrete chemical syntax yet lack 3D spatial understanding. Here we introduce SE3-BiLingoMol, an SE(3)-equivariant language model for pocket-aware 3D ligand de novo generation and fragment-guided optimization. Built upon Geometric Algebra Transformers and a fragment-aware SMILES representation, our model enables SE(3)-equivariant modeling of continuous 3D geometry while ensuring chemically valid molecular topologies. To counteract cumulative 3D conformational errors inherent to autoregressive generation, we developed a bidirectional attention-based self-refinement mechanism as a key architectural component of SE3-BiLingoMol. Our model achieves state-of-the-art performance in an in-silico evaluation across over 100 diverse targets. Critically, application of SE3-BiLingoMol led to the discovery of a novel tetracyclic HPK1 inhibitor showing potent in vitro activity and robust in vivo anti-tumor efficacy. This work demonstrates a powerful and practical generative AI framework for accelerating structure-based drug design.","manuscriptTitle":"An SE(3)-equivariant language model for pocket-aware 3D molecular generation enables discovery of potent HPK1 inhibitors","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-02-27 04:04:39","doi":"10.21203/rs.3.rs-8880270/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"3b196312-e0e7-47da-a361-1bade27267a7","owner":[],"postedDate":"February 27th, 2026","published":true,"recentEditorialEvents":[{"type":"decision","content":"Reject after peer review","date":"2026-05-12T08:04:18+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"This content is not available.","date":"2026-05-11T03:12:39+00:00","index":3,"fulltext":"This content is not available."}],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":63555669,"name":"Biological sciences/Drug discovery/Medicinal chemistry/Structure-based drug design"},{"id":63555670,"name":"Biological sciences/Chemical biology/Small molecules"},{"id":63555671,"name":"Biological sciences/Computational biology and bioinformatics/Machine learning"},{"id":63555672,"name":"Health sciences/Diseases/Cancer/Cancer therapy/Cancer immunotherapy"},{"id":63555673,"name":"Biological sciences/Drug discovery/Drug screening/Virtual screening"}],"tags":[],"updatedAt":"2026-05-12T08:07:38+00:00","versionOfRecord":[],"versionCreatedAt":"2026-02-27 04:04:39","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8880270","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8880270","identity":"rs-8880270","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}