{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"institution":[{"name":"bioRxiv"}],"indexed":{"date-parts":[[2026,7,16]],"date-time":"2026-07-16T14:12:43Z","timestamp":1784211163408,"version":"3.55.0"},"posted":{"date-parts":[[2026,2,23]]},"group-title":"Bioinformatics","reference-count":59,"publisher":"openRxiv","license":[{"start":{"date-parts":[[2026,2,23]],"date-time":"2026-02-23T00:00:00Z","timestamp":1771804800000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"DOI":"10.13039\/100000092","name":"National Library of Medicine","doi-asserted-by":"crossref","award":["R01LM014199"],"award-info":[{"award-number":["R01LM014199"]}],"id":[{"id":"10.13039\/100000092","id-type":"DOI","asserted-by":"crossref"}]},{"DOI":"10.13039\/100000002","name":"National Institutes of Health","doi-asserted-by":"crossref","award":["K99LM014428"],"award-info":[{"award-number":["K99LM014428"]}],"id":[{"id":"10.13039\/100000002","id-type":"DOI","asserted-by":"crossref"}]},{"name":"National Cancer Institute of the National Institutes of Health","award":["T32CA160056"],"award-info":[{"award-number":["T32CA160056"]}]},{"name":"National Cancer Institute of the National Institutes of Health","award":["R01CA297582"],"award-info":[{"award-number":["R01CA297582"]}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"accepted":{"date-parts":[[2026,7,13]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                <jats:p>Graph foundation models have emerged as powerful tools for drug repurposing by enabling the prediction of novel drug\u2013disease indications from large biomedical knowledge graphs. A representative example is TxGNN, which was previously developed and trained on PrimeKG, a comprehensive biomedical knowledge graph covering over 17,000 diseases. While TxGNN demonstrates strong performance, existing biomedical knowledge graphs largely lack fine-grained, cell-type-specific expression context. This limits their ability to capture disease mechanisms driven by dysregulated cellular programs, such as immune cell-specific pathways in autoimmune diseases. Moreover, prior evaluations typically test only randomly selected subsets of diseases, leaving many diseases unexamined and limiting conclusions about model performance across the full disease spectrum.<\/jats:p>\n                <jats:p>\n                  To address these limitations, we first update PrimeKG to\n                  <jats:bold>PrimeKG-U<\/jats:bold>\n                  by incorporating expanded and curated biomedical knowledge and then develop\n                  <jats:bold>TxGNN-U<\/jats:bold>\n                  as a stronger graph-based baseline. Building on this foundation, we introduce\n                  <jats:bold>CellAwareGNN<\/jats:bold>\n                  , a graph foundation model that integrates single-cell genomics into PrimeKG-U. We construct a single-cell-enhanced knowledge graph,\n                  <jats:bold>scPrimeKG<\/jats:bold>\n                  , by incorporating cell-type-specific gene expression signatures from the OneK1K dataset, expanding PrimeKG from approximately 8.1 million edges and 129k nodes to over 14 million edges and 147k nodes. CellAwareGNN is pre-trained on all relation types in scPrimeKG and evaluated on drug indication prediction with explicit coverage of all diseases in the knowledge graph.\n                <\/jats:p>\n                <jats:p>CellAwareGNN consistently outperforms TxGNN and TxGNN-U. For drug indication prediction, CellAwareGNN achieves an AUPRC of 0.826, representing a 1.2% improvement over TxGNN-U (0.816) and a 3.4% improvement over TxGNN (0.799). Notably, for autoimmune diseases, CellAwareGNN attains an AUPRC of 0.864, improving by 2.0% over TxGNN-U (0.847) and 6.0% over TxGNN (0.815). Importantly, CellAwareGNN prioritizes promising repurposing candidates, including Ocrelizumab for Pemphigus via CD20-expressing B cells, Methotrexate for Pemphigus through DHFR and ATIC activity in T and B cells, and Rosiglitazone for Rheumatoid Arthritis through PPAR-\u03b3 activation. These results demonstrate the value of incorporating cell-type-specific expression context to improve both predictive performance and biological interpretability in graph-based drug repurposing.<\/jats:p>\n                <jats:sec>\n                  <jats:title>CCS Concepts<\/jats:title>\n                  <jats:list list-type=\"bullet\">\n                    <jats:list-item>\n                      <jats:p>Applied computing \u2192 Health informatics; Bioinformatics;<\/jats:p>\n                    <\/jats:list-item>\n                    <jats:list-item>\n                      <jats:p>\n                        Computing methodologies \u2192 Knowledge representation and reasoning;\n                        <jats:italic>Neural networks<\/jats:italic>\n                        .\n                      <\/jats:p>\n                    <\/jats:list-item>\n                  <\/jats:list>\n                <\/jats:sec>\n                <jats:sec>\n                  <jats:title>ACM Reference Format<\/jats:title>\n                  <jats:p>\n                    Xinmeng Zhang, Eugene Jeong, Chao Yan, Yubo Feng, Linshuoshuo Lyu, Xingyi Guo, and You Chen. 2026. CellAwareGNN: Single-Cell Enhanced Knowledge Graph Foundation Model for Drug Indication Prediction. In\n                    <jats:italic>Proceedings of the 32nd ACM SIGKDD Conference on Knowledge Discovery and Data Mining V<\/jats:italic>\n                    .\n                    <jats:italic>2 (KDD 2026), August 9\u201313, 2026, Jeju Island, Republic of Korea<\/jats:italic>\n                    . ACM, New York, NY, USA, 12 pages.\n                    <jats:ext-link xmlns:xlink=\"http:\/\/www.w3.org\/1999\/xlink\" ext-link-type=\"uri\" xlink:href=\"https:\/\/doi.org\/10.1145\/3770855.3819012\">https:\/\/doi.org\/10.1145\/3770855.3819012<\/jats:ext-link>\n                  <\/jats:p>\n                <\/jats:sec>","DOI":"10.64898\/2026.02.20.707076","type":"posted-content","created":{"date-parts":[[2026,2,23]],"date-time":"2026-02-23T16:05:13Z","timestamp":1771862713000},"source":"Crossref","is-referenced-by-count":1,"title":["CellAwareGNN: Single-Cell Enhanced Knowledge Graph Foundation Model for Drug Indication Prediction"],"prefix":"10.64898","author":[{"ORCID":"https:\/\/orcid.org\/0000-0001-7876-0753","authenticated-orcid":false,"given":"Xinmeng","family":"Zhang","sequence":"first","affiliation":[{"name":"Department of Computer Science, Vanderbilt University, Nashville, TN, USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-8240-1198","authenticated-orcid":false,"given":"Eugene","family":"Jeong","sequence":"additional","affiliation":[{"name":"Department of Epidemiology, Vanderbilt University Medical Center, Nashville, TN, USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-6719-1388","authenticated-orcid":false,"given":"Chao","family":"Yan","sequence":"additional","affiliation":[{"name":"Department of Biomedical, Informatics, Vanderbilt University Medical Center, Nashville, TN, USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0003-9214-4991","authenticated-orcid":false,"given":"Yubo","family":"Feng","sequence":"additional","affiliation":[{"name":"Department of Computer Science, Vanderbilt University, Nashville, TN, USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Linshuoshuo","family":"Lyu","sequence":"additional","affiliation":[{"name":"Department of Epidemiology Vanderbilt University Nashville, TN, USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-5269-1294","authenticated-orcid":false,"given":"Xingyi","family":"Guo","sequence":"additional","affiliation":[{"name":"Department of Medicine, Vanderbilt University Medical, Center Nashville, TN, USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-8232-8840","authenticated-orcid":false,"given":"You","family":"Chen","sequence":"additional","affiliation":[{"name":"Department of Biomedical, Informatics, Vanderbilt University Medical Center, Nashville, TN, USA"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"54368","reference":[{"key":"2026071606450636000_2026.02.20.707076v2.1","unstructured":"2025. Table. Prednisolone-equivalent dosing, duration of action and anti-inflammatory activity for various systemic corticosteroid preparations | The Australian Immunisation Handbook. https:\/\/immunisationhandbook.health.gov.au\/resources\/tables\/table-prednisolone-equivalent-dosing-duration-of-action-and-anti-inflammatory-activity-for-various-systemic-corticosteroid-preparations"},{"key":"2026071606450636000_2026.02.20.707076v2.2","unstructured":"2026. Azathioprine therapy induces selective NK cell depletion and IFN-\u03b3 deficiency predisposing to herpesvirus reactivation - ClinicalKey. https:\/\/www.clinicalkey.com\/#!\/content\/playContent\/1-s2.0-S0091674922011873?returnurl=https:%2F%2Flinkinghub.elsevier.com%2Fretrieve%2Fpii%2FS0091674922011873%3Fshowall%3Dtrue&referrer="},{"key":"2026071606450636000_2026.02.20.707076v2.3","unstructured":"2026. From \u2018differential expression\u2019 to \u2018differential networking\u2019 \u2013 identification of dysfunctional regulatory networks in diseases - ClinicalKey. https:\/\/www-clinicalkey-com.proxy.library.vanderbilt.edu\/#!\/content\/playContent\/1-s2.0-S0168952510000879?returnurl=null&referrer=null"},{"key":"2026071606450636000_2026.02.20.707076v2.4","unstructured":"2026. Mellen Center Approach to Relapse Management in Multiple Sclerosis. https:\/\/my.clevelandclinic.org\/departments\/neurological\/depts\/multiple-sclerosis\/ms-approaches\/relapse-management-ms"},{"key":"2026071606450636000_2026.02.20.707076v2.5","unstructured":"2026. Mondo: integrating disease terminology across communities | Genetics | Oxford Academic. https:\/\/academic.oup.com\/genetics\/advance-article\/doi\/10.1093\/genetics\/iyaf215\/8276117"},{"key":"2026071606450636000_2026.02.20.707076v2.6","doi-asserted-by":"publisher","DOI":"10.1073\/pnas.2008523117"},{"key":"2026071606450636000_2026.02.20.707076v2.7","unstructured":"2026. VA.gov | Veterans Affairs. Accepted: 20240213. https:\/\/www.va.gov\/MS\/TREATING_MS\/Treatments_for_Multiple_Sclerosis_Relapses.asp?utm_source=openai"},{"key":"2026071606450636000_2026.02.20.707076v2.8","doi-asserted-by":"publisher","DOI":"10.1684\/ejd.2011.1611"},{"key":"2026071606450636000_2026.02.20.707076v2.9","doi-asserted-by":"publisher","DOI":"10.1016\/j.jdcr.2021.08.005"},{"key":"2026071606450636000_2026.02.20.707076v2.10","doi-asserted-by":"publisher","DOI":"10.1038\/s41597-023-02757-0"},{"key":"2026071606450636000_2026.02.20.707076v2.11","doi-asserted-by":"publisher","DOI":"10.1093\/rheumatology\/kead104.131"},{"key":"2026071606450636000_2026.02.20.707076v2.12","doi-asserted-by":"publisher","DOI":"10.1038\/s41597-023-01960-3"},{"key":"2026071606450636000_2026.02.20.707076v2.13","doi-asserted-by":"publisher","DOI":"10.1038\/clpt.2009.103"},{"key":"2026071606450636000_2026.02.20.707076v2.14","doi-asserted-by":"publisher","DOI":"10.1136\/ard-2022-223739"},{"key":"2026071606450636000_2026.02.20.707076v2.15","doi-asserted-by":"publisher","DOI":"10.1016\/j.compbiomed.2022.105992"},{"key":"2026071606450636000_2026.02.20.707076v2.16","doi-asserted-by":"publisher","DOI":"10.1038\/nature13835"},{"key":"2026071606450636000_2026.02.20.707076v2.17","doi-asserted-by":"publisher","DOI":"10.1038\/s41467-022-33026-0"},{"key":"2026071606450636000_2026.02.20.707076v2.18","doi-asserted-by":"publisher","DOI":"10.1038\/d41586-018-05267-x"},{"key":"2026071606450636000_2026.02.20.707076v2.19","doi-asserted-by":"publisher","DOI":"10.1002\/acr.24596"},{"key":"2026071606450636000_2026.02.20.707076v2.20","doi-asserted-by":"publisher","DOI":"10.1007\/s40257-019-00497-9"},{"key":"2026071606450636000_2026.02.20.707076v2.21","doi-asserted-by":"publisher","DOI":"10.1093\/nar\/gkad1005"},{"key":"2026071606450636000_2026.02.20.707076v2.22","unstructured":"GlaxoSmithKline. 2016. A Randomised, Double-blind, Placebo-controlled, Parallel Group Study to Investigate the Anti-inflammatory and Metabolic Effects of Rosiglitazone XR, 8mg Once Daily, in Subjects With Rheumatoid Arthritis. Clinical trial registration NCT00379600. clinicaltrials.gov. submitted: 2006\u201309-21. https:\/\/clinicaltrials.gov\/study\/NCT00379600"},{"key":"2026071606450636000_2026.02.20.707076v2.23","doi-asserted-by":"publisher","DOI":"10.1109\/JBHI.2020.2990797"},{"key":"2026071606450636000_2026.02.20.707076v2.24","doi-asserted-by":"publisher","DOI":"10.7554\/eLife.26726"},{"key":"2026071606450636000_2026.02.20.707076v2.25","doi-asserted-by":"publisher","DOI":"10.1038\/s41591-024-03233-x"},{"key":"2026071606450636000_2026.02.20.707076v2.26","doi-asserted-by":"publisher","DOI":"10.1146\/annurev-biodatasci-110723-024625"},{"key":"2026071606450636000_2026.02.20.707076v2.27","doi-asserted-by":"publisher","DOI":"10.1093\/nar\/gkad976"},{"key":"2026071606450636000_2026.02.20.707076v2.28","doi-asserted-by":"publisher","DOI":"10.1134\/S1068162023020139"},{"key":"2026071606450636000_2026.02.20.707076v2.29","doi-asserted-by":"publisher","DOI":"10.1007\/BF00542111"},{"key":"2026071606450636000_2026.02.20.707076v2.30","doi-asserted-by":"publisher","DOI":"10.1038\/s41592-024-02341-3"},{"key":"2026071606450636000_2026.02.20.707076v2.31","doi-asserted-by":"publisher","DOI":"10.1038\/s41401-022-00957-9"},{"key":"2026071606450636000_2026.02.20.707076v2.32","doi-asserted-by":"publisher","DOI":"10.1016\/j.jaad.2018.02.021"},{"key":"2026071606450636000_2026.02.20.707076v2.33","unstructured":"Engineering National Academies of Sciences, Health and Medicine Division, Board on Population Health and Public Health Practice, and Committee for the Assessment of NIH Research on Autoimmune Diseases. 2022. List of Autoimmune Diseases. In Enhancing NIH Research on Autoimmune Disease. National Academies Press (US). https:\/\/www.ncbi.nlm.nih.gov\/books\/NBK605867\/"},{"key":"2026071606450636000_2026.02.20.707076v2.34","doi-asserted-by":"publisher","DOI":"10.1109\/TCBB.2021.3060430"},{"key":"2026071606450636000_2026.02.20.707076v2.35","doi-asserted-by":"publisher","DOI":"10.1038\/nri.2017.76"},{"key":"2026071606450636000_2026.02.20.707076v2.36","doi-asserted-by":"publisher","DOI":"10.1093\/bib\/bbae461"},{"key":"2026071606450636000_2026.02.20.707076v2.37","doi-asserted-by":"publisher","DOI":"10.64898\/2026.01.05.697749"},{"key":"2026071606450636000_2026.02.20.707076v2.38","doi-asserted-by":"publisher","DOI":"10.1007\/s12016-017-8662-z"},{"key":"2026071606450636000_2026.02.20.707076v2.39","doi-asserted-by":"publisher","DOI":"10.3389\/fimmu.2019.01375"},{"key":"2026071606450636000_2026.02.20.707076v2.40","doi-asserted-by":"publisher","DOI":"10.1038\/nrd.2018.168"},{"key":"2026071606450636000_2026.02.20.707076v2.41","doi-asserted-by":"publisher","DOI":"10.1155\/2014\/147648"},{"key":"2026071606450636000_2026.02.20.707076v2.42","first-page":"579","article-title":"Developing a Search Engine for Precision Medicine","volume":"2020","year":"2020","journal-title":"AMIA Summits on Translational Science Proceedings"},{"key":"2026071606450636000_2026.02.20.707076v2.43","doi-asserted-by":"publisher","DOI":"10.1126\/sciimmunol.abf2846"},{"key":"2026071606450636000_2026.02.20.707076v2.44","doi-asserted-by":"publisher","DOI":"10.1038\/nbt0717-604"},{"issue":"2020","key":"2026071606450636000_2026.02.20.707076v2.45","first-page":"463","article-title":"A Literature-Based Knowledge Graph Embedding Method for Identifying Drug Repurposing Opportunities in Rare Diseases","volume":"25","year":"2020","journal-title":"Pacific Symposium on Biocomputing. Pacific Symposium on Biocomputing"},{"key":"2026071606450636000_2026.02.20.707076v2.46","doi-asserted-by":"publisher","DOI":"10.1111\/bjd.12474"},{"key":"2026071606450636000_2026.02.20.707076v2.47","doi-asserted-by":"publisher","DOI":"10.3390\/pharmaceutics14071464"},{"key":"2026071606450636000_2026.02.20.707076v2.48","doi-asserted-by":"publisher","DOI":"10.1093\/nar\/gkw993"},{"key":"2026071606450636000_2026.02.20.707076v2.49","doi-asserted-by":"publisher","DOI":"10.1038\/s41588-018-0089-9"},{"key":"2026071606450636000_2026.02.20.707076v2.50","doi-asserted-by":"publisher","DOI":"10.1080\/24699322.2020.1850866"},{"key":"2026071606450636000_2026.02.20.707076v2.51","doi-asserted-by":"publisher","DOI":"10.1172\/JCI182790"},{"key":"2026071606450636000_2026.02.20.707076v2.52","doi-asserted-by":"publisher","DOI":"10.1038\/s41597-020-0543-2"},{"key":"2026071606450636000_2026.02.20.707076v2.53","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.1412.6575"},{"key":"2026071606450636000_2026.02.20.707076v2.54","doi-asserted-by":"publisher","DOI":"10.1126\/science.abf3041"},{"key":"2026071606450636000_2026.02.20.707076v2.55","doi-asserted-by":"publisher","DOI":"10.1038\/s41590-019-0378-1"},{"key":"2026071606450636000_2026.02.20.707076v2.56","doi-asserted-by":"publisher","DOI":"10.1021\/acs.chemrev.5c00461"},{"key":"2026071606450636000_2026.02.20.707076v2.57","doi-asserted-by":"publisher","DOI":"10.1016\/j.jbi.2021.103696"},{"key":"2026071606450636000_2026.02.20.707076v2.58","doi-asserted-by":"publisher","DOI":"10.1093\/bib\/bbaf555"},{"key":"2026071606450636000_2026.02.20.707076v2.59","doi-asserted-by":"publisher","DOI":"10.1016\/S2589-7500(20)30192-8"}],"container-title":[],"original-title":[],"link":[{"URL":"https:\/\/syndication.highwire.org\/content\/doi\/10.64898\/2026.02.20.707076","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,7,16]],"date-time":"2026-07-16T13:45:17Z","timestamp":1784209517000},"score":1,"resource":{"primary":{"URL":"http:\/\/biorxiv.org\/lookup\/doi\/10.64898\/2026.02.20.707076"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,2,23]]},"references-count":59,"URL":"https:\/\/doi.org\/10.64898\/2026.02.20.707076","relation":{},"subject":[],"published":{"date-parts":[[2026,2,23]]},"subtype":"preprint"}}