Heterogeneous Medical Knowledge Graph Completion using GNN for inductive Link prediction and Medicine Recommendation

Heterogeneous Medical Knowledge Graph Completion using GNN for inductive Link prediction and Medicine Recommendation | 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 Heterogeneous Medical Knowledge Graph Completion using GNN for inductive Link prediction and Medicine Recommendation Sooraj Kannadiparambil, Keshab Nath, Christina Terese Joseph, and 2 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-5585570/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 In the biomedical realm, a Knowledge graph is something that combines different sources of information that are expert - driven into a graph, in which, nodes represent biomedical entities like diseases, genomes, symptoms, drugs etc. Relationship between various entities are represented by edges. This paper focuses on creating a heterogeneous medical knowledge graph and predicting missing relations between symptom to drug, medicine recommendation using Graph Neural Networks(GNN) and Graphsage networks. This paper aims to create a Medical Knowledge Graph based on multi modal data (text data, chemical structure, etc.) on diseases, symptoms and drugs, then use a Graph Neural Network and its variants to predict the missing links between symptoms and drugs and also use medicine recommendation and then compare with existing methods and highlight the advantages of proposed methodology including evaluating the model with existing algorithms. In this paper, we were able to construct a heterogeneous knowledge graph consisting of disease, symptoms and drugs as nodes. We were able to achieve an accuracy and prediction above 80 percentage, which was higher than current existing models including translational distance models, DistMult , etc. With the model, we were able to predict missing links between symptoms and drugs, which could be more helpful in research and also able to recommend medicine, given a set of symptoms. Physical sciences/Energy science and technology Physical sciences/Engineering Biomedical Knowledge Graph Fisher exact Test Heterogenous Graph Neural Networks Sage convolution networks Medicine recommendation inductive learning Link prediction Full Text Additional Declarations No competing interests reported. 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-5585570","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":441296292,"identity":"58bd103a-44a2-40bb-a7b2-687dbeae37a8","order_by":0,"name":"Sooraj Kannadiparambil","email":"","orcid":"","institution":"Indian Institute of Information Technology, Kottayam","correspondingAuthor":false,"prefix":"","firstName":"Sooraj","middleName":"","lastName":"Kannadiparambil","suffix":""},{"id":441296293,"identity":"9dd0865e-8335-4f98-8582-5ed7d429b353","order_by":1,"name":"Keshab Nath","email":"","orcid":"","institution":"Bhattadev University Bajali","correspondingAuthor":false,"prefix":"","firstName":"Keshab","middleName":"","lastName":"Nath","suffix":""},{"id":441296294,"identity":"a78c2ea9-90e2-4fcc-bc77-31d923ddec31","order_by":2,"name":"Christina Terese Joseph","email":"","orcid":"","institution":"Indian Institute of Information Technology, Kottayam","correspondingAuthor":false,"prefix":"","firstName":"Christina","middleName":"Terese","lastName":"Joseph","suffix":""},{"id":441296295,"identity":"2dd762c3-0add-4e79-8093-d1f8af7325e1","order_by":3,"name":"John Paul Martin","email":"","orcid":"","institution":"Indian Institute of Information Technology, Kottayam","correspondingAuthor":false,"prefix":"","firstName":"John","middleName":"Paul","lastName":"Martin","suffix":""},{"id":441296296,"identity":"4f3ec9fb-4dd8-43ac-a0ee-b91c9a96c6d7","order_by":4,"name":"Sk Mahmudul Hassan","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABFElEQVRIiWNgGAWjYFACHiBmY5BhkIBw5UDEAQYGCyCVgFcLD0yLMVSLBPFaEhsgNG4tuu1nj27mKTvMwyDd/Pjjj5o76Wvbzx48XFAjwcDPnmOATYvZmby02zzngFpkjplJ8xx7lrvtTF7C4RnHJBgke95g13Igx+w2bxtQi0SCGTMD2+HcbQdyDA7zsEkwGNzAYcv5NzAt6Z8//vh3OB0oAtTyT4LBHpeWG3BbcgwkgIwEoIjBYd42oC0SuLS8Mbs551w60CU5ZdK8fYcNt90A2jKzT4JH4syzAuwOywHqKrOW45dI3/zxx7fD8kAR488F32zk+NuTN2ANZRhgQ+YwM0DiiwTATJryUTAKRsEoGOYAAIEyY9fA1maZAAAAAElFTkSuQmCC","orcid":"","institution":"Manipal Academy of Higher Education","correspondingAuthor":true,"prefix":"","firstName":"Sk","middleName":"Mahmudul","lastName":"Hassan","suffix":""}],"badges":[],"createdAt":"2024-12-05 09:23:15","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-5585570/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-5585570/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":102962236,"identity":"42c143cd-bf22-40b1-93ed-f26759cb7b7b","added_by":"auto","created_at":"2026-02-19 04:05:56","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1038858,"visible":true,"origin":"","legend":"","description":"","filename":"HeterogenousMedicalPaper.pdf","url":"https://assets-eu.researchsquare.com/files/rs-5585570/v1_covered_2e84bcd4-1e16-4774-ab3d-624ea451ddfa.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Heterogeneous Medical Knowledge Graph Completion using GNN for inductive Link prediction and Medicine Recommendation","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"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":"Biomedical Knowledge Graph, Fisher exact Test, Heterogenous Graph Neural Networks, Sage convolution networks, Medicine recommendation, inductive learning, Link prediction","lastPublishedDoi":"10.21203/rs.3.rs-5585570/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-5585570/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"In the biomedical realm, a Knowledge graph is something that combines different sources of information that are expert - driven into a graph, in which, nodes represent biomedical entities like diseases, genomes, symptoms, drugs etc. Relationship between various entities are represented by edges. This paper focuses on creating a heterogeneous medical knowledge graph and predicting missing relations between symptom to drug, medicine recommendation using Graph Neural Networks(GNN) and Graphsage networks. This paper aims to create a Medical Knowledge Graph based on multi modal data (text data, chemical structure, etc.) on diseases, symptoms and drugs, then use a Graph Neural Network and its variants to predict the missing links between symptoms and drugs and also use medicine recommendation and then compare with existing methods and highlight the advantages of proposed methodology including evaluating the model with existing algorithms. In this paper, we were able to construct a heterogeneous knowledge graph consisting of disease, symptoms and drugs as nodes. We were able to achieve an accuracy and prediction above 80 percentage, which was higher than current existing models including translational distance models, DistMult , etc. With the model, we were able to predict missing links between symptoms and drugs, which could be more helpful in research and also able to recommend medicine, given a set of symptoms.\n","manuscriptTitle":"Heterogeneous Medical Knowledge Graph Completion using GNN for inductive Link prediction and Medicine Recommendation","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-04-11 06:03:25","doi":"10.21203/rs.3.rs-5585570/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":"188ea9d1-5b56-4f8e-ad80-7a3f7c3f0b7a","owner":[],"postedDate":"April 11th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":46990081,"name":"Physical sciences/Energy science and technology"},{"id":46990082,"name":"Physical sciences/Engineering"}],"tags":[],"updatedAt":"2026-02-11T13:42:22+00:00","versionOfRecord":[],"versionCreatedAt":"2025-04-11 06:03:25","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-5585570","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-5585570","identity":"rs-5585570","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}