UAlign: Pushing the Limit of Template-free Retrosynthesis Prediction with Unsupervised SMILES Alignment

UAlign: Pushing the Limit of Template-free Retrosynthesis Prediction with Unsupervised SMILES Alignment | 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 Research Article UAlign: Pushing the Limit of Template-free Retrosynthesis Prediction with Unsupervised SMILES Alignment Kaipeng Zeng, Bo Yang, Xin Zhao, Yu Zhang, Fan Nie, Xiaokang Yang, and 2 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-4292195/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 15 Jul, 2024 Read the published version in Journal of Cheminformatics → Version 1 posted 10 You are reading this latest preprint version Abstract \textbf{Motivation} Retrosynthesis planning poses a formidable challenge in the organic chemical industry, particularly in pharmaceuticals. Single-step retrosynthesis prediction, a crucial step in the planning process, has witnessed a surge in interest in recent years due to advancements in AI for science. Various deep learning-based methods have been proposed for this task in recent years, incorporating diverse levels of additional chemical knowledge dependency. \vspace{2mm}\ \textbf{Results} This paper introduces {\modelname}, a template-free graph-to-sequence pipeline for retrosynthesis prediction. By combining graph neural networks and Transformers, our method can more effectively leverage the inherent graph structure of molecules. Based on the fact that the majority of molecule structures remain unchanged during a chemical reaction, we propose a simple yet effective SMILES alignment technique to facilitate the reuse of unchanged structures for reactant generation. Extensive experiments show that our method substantially outperforms state-of-the-art template-free and semi-template-based approaches. Importantly, our template-free method achieves effectiveness comparable to, or even surpasses, established powerful template-based methods. \vspace{2mm}\ \textbf{Scientific contribution} We present a novel graph-to-sequence template-free retrosynthesis prediction pipeline that overcomes the limitations of Transformer-based methods in molecular representation learning and insufficient utilization of chemical information. We propose an unsupervised learning mechanism for establishing product-atom correspondence with reactant SMILES tokens, achieving even better results than supervised SMILES alignment methods. Extensive experiments demonstrate that {\modelname} significantly outperforms state-of-the-art template-free methods and rivals or surpasses template-based approaches, with up to 5% (top-5) and 5.4% (top-10) increased accuracy over the strongest baseline. Template-Free Retrosynthesis Deep Learning Chemical Reactions Single-step Retrosynthesis Prediction Full Text Additional Declarations No competing interests reported. Supplementary Files SIJcheminfoUAlignzengkaipeng.zip Cite Share Download PDF Status: Published Journal Publication published 15 Jul, 2024 Read the published version in Journal of Cheminformatics → Version 1 posted Editorial decision: Revision requested 07 Jun, 2024 Reviews received at journal 03 Jun, 2024 Reviews received at journal 30 May, 2024 Reviewers agreed at journal 28 May, 2024 Reviewers agreed at journal 26 May, 2024 Reviewers agreed at journal 21 May, 2024 Reviewers invited by journal 10 May, 2024 Submission checks completed at journal 20 Apr, 2024 Editor assigned by journal 20 Apr, 2024 First submitted to journal 19 Apr, 2024 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-4292195","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":294465502,"identity":"f5f0abcb-8b57-4ada-8dd9-c09071f3eb95","order_by":0,"name":"Kaipeng Zeng","email":"","orcid":"","institution":"Shanghai Jiao Tong University","correspondingAuthor":false,"prefix":"","firstName":"Kaipeng","middleName":"","lastName":"Zeng","suffix":""},{"id":294465503,"identity":"76a8d1a5-171f-4c3d-87d5-012b5cdb37be","order_by":1,"name":"Bo Yang","email":"","orcid":"","institution":"Shanghai Jiao Tong University","correspondingAuthor":false,"prefix":"","firstName":"Bo","middleName":"","lastName":"Yang","suffix":""},{"id":294465504,"identity":"0c7e2281-64c8-42fd-9d3d-c81d7a7ab53a","order_by":2,"name":"Xin Zhao","email":"","orcid":"","institution":"Shanghai Jiao Tong University","correspondingAuthor":false,"prefix":"","firstName":"Xin","middleName":"","lastName":"Zhao","suffix":""},{"id":294465505,"identity":"19ef6fd3-6d5b-4eb8-82aa-a4921bef4433","order_by":3,"name":"Yu Zhang","email":"","orcid":"","institution":"Shanghai Jiao Tong University","correspondingAuthor":false,"prefix":"","firstName":"Yu","middleName":"","lastName":"Zhang","suffix":""},{"id":294465506,"identity":"7c427f3b-ba39-4c67-840b-fd9b08c0a28a","order_by":4,"name":"Fan Nie","email":"","orcid":"","institution":"Shanghai Jiao Tong University","correspondingAuthor":false,"prefix":"","firstName":"Fan","middleName":"","lastName":"Nie","suffix":""},{"id":294465507,"identity":"3832e95d-dec6-4170-b8b5-37a772c82c5b","order_by":5,"name":"Xiaokang Yang","email":"","orcid":"","institution":"Shanghai Jiao Tong University","correspondingAuthor":false,"prefix":"","firstName":"Xiaokang","middleName":"","lastName":"Yang","suffix":""},{"id":294465508,"identity":"66f82028-b35d-481d-ae7a-d3b069c6ac5c","order_by":6,"name":"Yaohui Jin","email":"","orcid":"","institution":"Shanghai Jiao Tong University","correspondingAuthor":false,"prefix":"","firstName":"Yaohui","middleName":"","lastName":"Jin","suffix":""},{"id":294465509,"identity":"c3d675fc-e036-4ed0-955f-e51ad396523e","order_by":7,"name":"Yanyan Xu","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA5klEQVRIiWNgGAWjYDADfgaGBBDN2EC0FskGhsQG0rQYHICoJqzF4PjZYxI/dxxO3Hz+wPPHPAw2shsOMD97gFfLmbw0yd4zaYnbbiQkNvMwpBlvOMBmboDfPTlmErxtNkAtDCAthxM3HOBhk8Cr5fwbM8m/bRKJm/sPgLT8J0LLjRwzaZAtGxjADjtAWIvkjTfG1rJtacYzgH6ZOccg2XjmYTYzvFr4zucY3nzbdli2v/9Mwoc3FXayfcebn+HVonCAgQWqgCcB6E4gzYxPPRDINzAwf4Aw2Q8QUDsKRsEoGAUjFQAAEzRPEOxTw0QAAAAASUVORK5CYII=","orcid":"","institution":"Shanghai Jiao Tong University","correspondingAuthor":true,"prefix":"","firstName":"Yanyan","middleName":"","lastName":"Xu","suffix":""}],"badges":[],"createdAt":"2024-04-19 09:23:25","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-4292195/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-4292195/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1186/s13321-024-00877-2","type":"published","date":"2024-07-15T16:12:58+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":61595213,"identity":"6288f1c2-47fc-4c31-b380-5297ffc008ba","added_by":"auto","created_at":"2024-08-01 17:21:00","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":829974,"visible":true,"origin":"","legend":"","description":"","filename":"JChemInfoUalignzengkaipeng.pdf","url":"https://assets-eu.researchsquare.com/files/rs-4292195/v1_covered_1ad4478f-fda8-4df0-8f9f-cadf1af36b06.pdf"},{"id":55258020,"identity":"334bf99b-fba2-4cba-9c71-11a761dde8d1","added_by":"auto","created_at":"2024-04-24 21:00:11","extension":"zip","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":1764573,"visible":true,"origin":"","legend":"","description":"","filename":"SIJcheminfoUAlignzengkaipeng.zip","url":"https://assets-eu.researchsquare.com/files/rs-4292195/v1/234995bbe4d3b3130966cc71.zip"}],"financialInterests":"No competing interests reported.","formattedTitle":"UAlign: Pushing the Limit of Template-free Retrosynthesis Prediction with Unsupervised SMILES Alignment","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":"journal-of-cheminformatics","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"chin","sideBox":"Learn more about [Journal of Cheminformatics](https://jcheminf.biomedcentral.com/)","snPcode":"13321","submissionUrl":"https://submission.nature.com/new-submission/13321/3","title":"Journal of Cheminformatics","twitterHandle":"@jcheminf","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"BMC/SO AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Template-Free Retrosynthesis, Deep Learning, Chemical Reactions, Single-step Retrosynthesis Prediction","lastPublishedDoi":"10.21203/rs.3.rs-4292195/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-4292195/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\\textbf{Motivation} Retrosynthesis planning poses a formidable challenge in the organic chemical industry, particularly in pharmaceuticals. Single-step retrosynthesis prediction, a crucial step in the planning process, has witnessed a surge in interest in recent years due to advancements in AI for science. Various deep learning-based methods have been proposed for this task in recent years, incorporating diverse levels of additional chemical knowledge dependency. \\vspace{2mm}\\\\ \\textbf{Results} This paper introduces {\\modelname}, a template-free graph-to-sequence pipeline for retrosynthesis prediction. By combining graph neural networks and Transformers, our method can more effectively leverage the inherent graph structure of molecules. Based on the fact that the majority of molecule structures remain unchanged during a chemical reaction, we propose a simple yet effective SMILES alignment technique to facilitate the reuse of unchanged structures for reactant generation. Extensive experiments show that our method substantially outperforms state-of-the-art template-free and semi-template-based approaches. Importantly, our template-free method achieves effectiveness comparable to, or even surpasses, established powerful template-based methods. \\vspace{2mm}\\\\ \\textbf{Scientific contribution} We present a novel graph-to-sequence template-free retrosynthesis prediction pipeline that overcomes the limitations of Transformer-based methods in molecular representation learning and insufficient utilization of chemical information. We propose an unsupervised learning mechanism for establishing product-atom correspondence with reactant SMILES tokens, achieving even better results than supervised SMILES alignment methods. Extensive experiments demonstrate that {\\modelname} significantly outperforms state-of-the-art template-free methods and rivals or surpasses template-based approaches, with up to 5\\% (top-5) and 5.4\\% (top-10) increased accuracy over the strongest baseline.","manuscriptTitle":"UAlign: Pushing the Limit of Template-free Retrosynthesis Prediction with Unsupervised SMILES Alignment","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-04-24 21:00:06","doi":"10.21203/rs.3.rs-4292195/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2024-06-08T02:15:24+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2024-06-03T14:03:10+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2024-05-30T22:41:46+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"250953136961839462477801665252514065721","date":"2024-05-28T05:36:54+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"134658507450331018139166449054407918462","date":"2024-05-26T22:28:21+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"161337746717852870263224704546349480045","date":"2024-05-21T07:58:07+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2024-05-10T13:38:56+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2024-04-20T11:22:09+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2024-04-20T11:22:09+00:00","index":"","fulltext":""},{"type":"submitted","content":"Journal of Cheminformatics","date":"2024-04-19T09:22:09+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"journal-of-cheminformatics","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"chin","sideBox":"Learn more about [Journal of Cheminformatics](https://jcheminf.biomedcentral.com/)","snPcode":"13321","submissionUrl":"https://submission.nature.com/new-submission/13321/3","title":"Journal of Cheminformatics","twitterHandle":"@jcheminf","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"BMC/SO AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"c280fc34-0c46-4c3b-b375-0437d142bd93","owner":[],"postedDate":"April 24th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[],"tags":[],"updatedAt":"2024-08-01T16:18:27+00:00","versionOfRecord":{"articleIdentity":"rs-4292195","link":"https://doi.org/10.1186/s13321-024-00877-2","journal":{"identity":"journal-of-cheminformatics","isVorOnly":false,"title":"Journal of Cheminformatics"},"publishedOn":"2024-07-15 16:12:58","publishedOnDateReadable":"July 15th, 2024"},"versionCreatedAt":"2024-04-24 21:00:06","video":"","vorDoi":"10.1186/s13321-024-00877-2","vorDoiUrl":"https://doi.org/10.1186/s13321-024-00877-2","workflowStages":[]},"version":"v1","identity":"rs-4292195","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-4292195","identity":"rs-4292195","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}