GSMFormer-PPI: Predicting Protein-Protein Interactions with Multimodal Graph, Surface, and Language Representations

GSMFormer-PPI: Predicting Protein-Protein Interactions with Multimodal Graph, Surface, and Language Representations | 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 GSMFormer-PPI: Predicting Protein-Protein Interactions with Multimodal Graph, Surface, and Language Representations David Arteaga, Nikita Chervov, Maria Poptsova This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7623993/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 06 Jan, 2026 Read the published version in Scientific Reports → Version 1 posted 10 You are reading this latest preprint version Abstract Accurate prediction of protein-protein interactions (PPIs) is fundamental to understanding biological processes and disease mechanisms. While deep learning offers a powerful alternative to costly experimental methods, existing approaches often overlook critical protein surface information and rely on simplistic feature fusion techniques, limiting their performance. To address this, we introduce GSMFormer-PPI, a novel multimodal framework that integrates protein molecular surface features, 3D structural graphs, and residue-level sequence embeddings. Our architecture employs geometric deep learning (MaSIF) to extract physicochemical surface descriptors, graph convolutional networks to process structural context, and a transformer encoder with linear projectors to learn complex, cross-modal interactions beyond simple concatenation. Evaluated on a curated dataset from PINDER, GSMFormer-PPI outperforms state-of-the-art graph-based and multimodal models. Ablation studies confirm the critical contribution of surface features and our advanced fusion strategy to the model's superior predictive power. This work demonstrates that the integrative analysis of surface, structure, and sequence data is a vital and promising direction for advancing PPI prediction. Biological sciences/Computational biology and bioinformatics Physical sciences/Mathematics and computing Full Text Additional Declarations No competing interests reported. Cite Share Download PDF Status: Published Journal Publication published 06 Jan, 2026 Read the published version in Scientific Reports → Version 1 posted Editorial decision: Revision requested 24 Oct, 2025 Reviews received at journal 19 Oct, 2025 Reviews received at journal 08 Oct, 2025 Reviewers agreed at journal 26 Sep, 2025 Reviewers agreed at journal 24 Sep, 2025 Reviewers invited by journal 24 Sep, 2025 Editor invited by journal 19 Sep, 2025 Editor assigned by journal 17 Sep, 2025 Submission checks completed at journal 16 Sep, 2025 First submitted to journal 15 Sep, 2025 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-7623993","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":524606172,"identity":"7c25d8ba-af76-4225-8307-0c3ab24059de","order_by":0,"name":"David Arteaga","email":"","orcid":"","institution":"National Research University Higher School of Economics","correspondingAuthor":false,"prefix":"","firstName":"David","middleName":"","lastName":"Arteaga","suffix":""},{"id":524606173,"identity":"f0bce3f9-c106-494b-adbd-a4f7b1ccc266","order_by":1,"name":"Nikita Chervov","email":"","orcid":"","institution":"National Research University Higher School of Economics","correspondingAuthor":false,"prefix":"","firstName":"Nikita","middleName":"","lastName":"Chervov","suffix":""},{"id":524606174,"identity":"a95bcbc2-ca90-4025-b64b-091408536470","order_by":2,"name":"Maria Poptsova","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAsUlEQVRIiWNgGAWjYFAD9h5SVB8AETxnSNYikUOkat3Zzccef/hjk9g/8+3hDx8Y7iU2ENJidudYusHBtrTEGbfz0iRnMBQToeVGjpnEwYbDxgy3c8yYeRgSiNRy4M9hY/mbZ4w//yFeC9thOYMbPAbSDERpAfnlbFuanOGZHDPJHoMEY8Jabjcfe1Dxx4ZH7vgZ4w8/KhJkCWphkGBgQ+IZEFSPoWUUjIJRMApGARYAAPhZQRnSYojKAAAAAElFTkSuQmCC","orcid":"","institution":"National Research University Higher School of Economics","correspondingAuthor":true,"prefix":"","firstName":"Maria","middleName":"","lastName":"Poptsova","suffix":""}],"badges":[],"createdAt":"2025-09-15 21:08:12","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7623993/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7623993/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1038/s41598-025-34758-x","type":"published","date":"2026-01-06T15:57:19+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":92893583,"identity":"f70a2786-1558-4c47-a005-ef6843ec9186","added_by":"auto","created_at":"2025-10-06 18:36:44","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":5514868,"visible":true,"origin":"","legend":"","description":"","filename":"GSMFormerPPI150925.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7623993/v1/714fe2727b679ef18d002975.pdf"},{"id":92893582,"identity":"0ff8f998-3f2b-4465-896d-a85657fb085d","added_by":"auto","created_at":"2025-10-06 18:36:44","extension":"json","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":5051,"visible":true,"origin":"","legend":"","description":"","filename":"6df335c6a9df4787b133726b906957c7.json","url":"https://assets-eu.researchsquare.com/files/rs-7623993/v1/fc9b78bde090a74fcc854459.json"},{"id":100069359,"identity":"df225e11-d5f0-485f-bf3e-dd3e3970dfc3","added_by":"auto","created_at":"2026-01-12 16:13:23","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1945862,"visible":true,"origin":"","legend":"","description":"","filename":"GSMFormerPPI150925.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7623993/v1_covered_eb09aa4b-ec61-49c8-b46f-d91ed4db2f3a.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"GSMFormer-PPI: Predicting Protein-Protein Interactions with Multimodal Graph, Surface, and Language Representations","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":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-7623993/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7623993/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"Accurate prediction of protein-protein interactions (PPIs) is fundamental to understanding biological processes and disease mechanisms. While deep learning offers a powerful alternative to costly experimental methods, existing approaches often overlook critical protein surface information and rely on simplistic feature fusion techniques, limiting their performance. To address this, we introduce GSMFormer-PPI, a novel multimodal framework that integrates protein molecular surface features, 3D structural graphs, and residue-level sequence embeddings. Our architecture employs geometric deep learning (MaSIF) to extract physicochemical surface descriptors, graph convolutional networks to process structural context, and a transformer encoder with linear projectors to learn complex, cross-modal interactions beyond simple concatenation. Evaluated on a curated dataset from PINDER, GSMFormer-PPI outperforms state-of-the-art graph-based and multimodal models. Ablation studies confirm the critical contribution of surface features and our advanced fusion strategy to the model's superior predictive power. This work demonstrates that the integrative analysis of surface, structure, and sequence data is a vital and promising direction for advancing PPI prediction.","manuscriptTitle":"GSMFormer-PPI: Predicting Protein-Protein Interactions with Multimodal Graph, Surface, and Language Representations","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-10-06 18:36:39","doi":"10.21203/rs.3.rs-7623993/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2025-10-24T04:14:55+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-10-19T21:32:37+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-10-08T12:27:51+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"36195664438349353800998329753927333630","date":"2025-09-26T16:10:13+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"252544369986422181305601738722125663474","date":"2025-09-24T11:23:11+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-09-24T09:27:57+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2025-09-19T05:36:36+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-09-18T01:56:58+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-09-17T01:34:32+00:00","index":"","fulltext":""},{"type":"submitted","content":"Scientific Reports","date":"2025-09-15T20:53:31+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"8bf69395-5156-43fc-ba7d-8331052bda99","owner":[],"postedDate":"October 6th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[{"id":55764152,"name":"Biological sciences/Computational biology and bioinformatics"},{"id":55764153,"name":"Physical sciences/Mathematics and computing"}],"tags":[],"updatedAt":"2026-01-12T16:04:50+00:00","versionOfRecord":{"articleIdentity":"rs-7623993","link":"https://doi.org/10.1038/s41598-025-34758-x","journal":{"identity":"scientific-reports","isVorOnly":false,"title":"Scientific Reports"},"publishedOn":"2026-01-06 15:57:19","publishedOnDateReadable":"January 6th, 2026"},"versionCreatedAt":"2025-10-06 18:36:39","video":"","vorDoi":"10.1038/s41598-025-34758-x","vorDoiUrl":"https://doi.org/10.1038/s41598-025-34758-x","workflowStages":[]},"version":"v1","identity":"rs-7623993","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7623993","identity":"rs-7623993","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}