{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"institution":[{"name":"medRxiv"}],"indexed":{"date-parts":[[2026,6,24]],"date-time":"2026-06-24T10:14:51Z","timestamp":1782296091866,"version":"3.54.5"},"posted":{"date-parts":[[2023,3,20]]},"group-title":"Health Informatics","reference-count":60,"publisher":"openRxiv","license":[{"start":{"date-parts":[[2023,3,20]],"date-time":"2023-03-20T00:00:00Z","timestamp":1679270400000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"accepted":{"date-parts":[[2024,8,7]]},"abstract":"<jats:p>\n                  Drug repurposing \u2013 identifying new therapeutic uses for approved drugs \u2013 is often serendipitous and opportunistic, expanding the use of drugs for new diseases. The clinical utility of drug repurposing AI models remains limited because the models focus narrowly on diseases for which some drugs already exist. Here, we introduce T\n                  <jats:sc>x<\/jats:sc>\n                  GNN, a graph foundation model for zero-shot drug repurposing, identifying therapeutic candidates even for diseases with limited treatment options or no existing drugs. Trained on a medical knowledge graph, T\n                  <jats:sc>x<\/jats:sc>\n                  GNN utilizes a graph neural network and metric-learning module to rank drugs as potential indications and contraindications across 17,080 diseases. When benchmarked against eight methods, T\n                  <jats:sc>x<\/jats:sc>\n                  GNN improves prediction accuracy for indications by 49.2% and contraindications by 35.1% under stringent zero-shot evaluation. To facilitate model interpretation, T\n                  <jats:sc>x<\/jats:sc>\n                  GNN\u2019s Explainer module offers transparent insights into multi-hop medical knowledge paths that form T\n                  <jats:sc>x<\/jats:sc>\n                  GNN\u2019s predictive rationales. Human evaluation of T\n                  <jats:sc>x<\/jats:sc>\n                  GNN\u2019s Explainer showed that T\n                  <jats:sc>x<\/jats:sc>\n                  GNN\u2019s predictions and explanations perform encouragingly on multiple axes of performance beyond accuracy. Many of T\n                  <jats:sc>x<\/jats:sc>\n                  GNN\u2019s novel predictions align with off-label prescriptions clinicians make in a large healthcare system. T\n                  <jats:sc>x<\/jats:sc>\n                  GNN\u2019s drug repurposing predictions are accurate, consistent with off-label drug use, and can be investigated by human experts through multi-hop interpretable rationales.\n                <\/jats:p>","DOI":"10.1101\/2023.03.19.23287458","type":"posted-content","created":{"date-parts":[[2023,3,20]],"date-time":"2023-03-20T15:15:12Z","timestamp":1679325312000},"source":"Crossref","is-referenced-by-count":14,"title":["A foundation model for clinician-centered drug repurposing"],"prefix":"10.64898","author":[{"ORCID":"https:\/\/orcid.org\/0000-0001-6693-8390","authenticated-orcid":false,"given":"Kexin","family":"Huang","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-1097-803X","authenticated-orcid":false,"given":"Payal","family":"Chandak","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-1728-4102","authenticated-orcid":false,"given":"Qianwen","family":"Wang","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-8783-7791","authenticated-orcid":false,"given":"Shreyas","family":"Havaldar","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-3343-744X","authenticated-orcid":false,"given":"Akhil","family":"Vaid","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-5411-923X","authenticated-orcid":false,"given":"Jure","family":"Leskovec","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-6319-4314","authenticated-orcid":false,"given":"Girish","family":"Nadkarni","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-4515-8090","authenticated-orcid":false,"given":"Benjamin S.","family":"Glicksberg","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-0327-8297","authenticated-orcid":false,"given":"Nils","family":"Gehlenborg","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-8530-7228","authenticated-orcid":false,"given":"Marinka","family":"Zitnik","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"54368","reference":[{"key":"2024080900100562000_2023.03.19.23287458v4.1","doi-asserted-by":"crossref","first-page":"165","DOI":"10.1001\/jamaneurol.2020.4152","article-title":"Burden of neurological disorders across the us from 1990-2017: a global burden of disease study","volume":"78","year":"2021","journal-title":"JAMA Neurology"},{"key":"2024080900100562000_2023.03.19.23287458v4.2","doi-asserted-by":"crossref","first-page":"157","DOI":"10.1038\/449157a","article-title":"Neglected diseases","volume":"449","year":"2007","journal-title":"Nature"},{"key":"2024080900100562000_2023.03.19.23287458v4.3","unstructured":"U.S. Food and Drug Administration. Rare Disease Day 2021. https:\/\/www.fda.gov\/news-events\/fda-voices\/rare-disease-day-2021-fda-shows-sustained-support-rare-disease-product-development-during-public (2023)."},{"key":"2024080900100562000_2023.03.19.23287458v4.4","doi-asserted-by":"publisher","DOI":"10.1038\/nrd.2018.168"},{"key":"2024080900100562000_2023.03.19.23287458v4.5","doi-asserted-by":"crossref","first-page":"751","DOI":"10.1038\/s41569-022-00717-6","article-title":"Repurposing drugs to treat cardiovascular disease in the era of precision medicine","volume":"19","year":"2022","journal-title":"Nature Reviews Cardiology"},{"key":"2024080900100562000_2023.03.19.23287458v4.6","doi-asserted-by":"crossref","unstructured":"Sahragardjoonegani, B. , Beall, R. F. , Kesselheim, A. S. & Hollis, A. Repurposing existing drugs for new uses: a cohort study of the frequency of FDA-granted new indication exclusivities since 1997. Journal of Pharmaceutical Policy and Practice 14 (2021).","DOI":"10.1186\/s40545-020-00282-8"},{"key":"2024080900100562000_2023.03.19.23287458v4.7","doi-asserted-by":"publisher","DOI":"10.1093\/bib\/bbr021"},{"key":"2024080900100562000_2023.03.19.23287458v4.8","doi-asserted-by":"publisher","DOI":"10.1111\/jphp.13273"},{"key":"2024080900100562000_2023.03.19.23287458v4.9","doi-asserted-by":"crossref","first-page":"67","DOI":"10.1038\/s41597-023-01960-3","article-title":"Building a knowledge graph to enable precision medicine","volume":"10","year":"2023","journal-title":"Scientific Data"},{"key":"2024080900100562000_2023.03.19.23287458v4.10","doi-asserted-by":"crossref","unstructured":"Menche, J. et al. Uncovering disease-disease relationships through the incomplete interactome. Science 347 (2015).","DOI":"10.1126\/science.1257601"},{"key":"2024080900100562000_2023.03.19.23287458v4.11","doi-asserted-by":"publisher","DOI":"10.1073\/pnas.1818013116"},{"key":"2024080900100562000_2023.03.19.23287458v4.12","first-page":"1","article-title":"Identification of disease treatment mechanisms through the multiscale interactome","volume":"12","year":"2021","journal-title":"Nature Communications"},{"key":"2024080900100562000_2023.03.19.23287458v4.13","doi-asserted-by":"publisher","DOI":"10.1073\/pnas.0701361104"},{"key":"2024080900100562000_2023.03.19.23287458v4.14","doi-asserted-by":"publisher","DOI":"10.1038\/nrg2918"},{"key":"2024080900100562000_2023.03.19.23287458v4.15","unstructured":"Li, M. M. , Huang, K. & Zitnik, M. Graph representation learning in biomedicine and healthcare. Nature Biomedical Engineering 1\u201317 (2022)."},{"key":"2024080900100562000_2023.03.19.23287458v4.16","doi-asserted-by":"crossref","unstructured":"Gysi, D. M. et al. Network medicine framework for identifying drug-repurposing opportunities for covid-19. Proceedings of the National Academy of Sciences 118 (2021).","DOI":"10.1073\/pnas.2025581118"},{"key":"2024080900100562000_2023.03.19.23287458v4.17","doi-asserted-by":"publisher","DOI":"10.1371\/journal.pone.0076339"},{"key":"2024080900100562000_2023.03.19.23287458v4.18","doi-asserted-by":"publisher","DOI":"10.1093\/bioinformatics\/bty294"},{"key":"2024080900100562000_2023.03.19.23287458v4.19","first-page":"1","article-title":"Network-based in silico drug efficacy screening","volume":"7","year":"2016","journal-title":"Nature Communications"},{"key":"2024080900100562000_2023.03.19.23287458v4.20","first-page":"1","article-title":"Network-based prediction of drug combinations","volume":"10","year":"2019","journal-title":"Nature Communications"},{"key":"2024080900100562000_2023.03.19.23287458v4.21","first-page":"1","article-title":"A comprehensive study of the rare diseases and conditions targeted by orphan drug designations and approvals over the forty years of the orphan drug act","volume":"18","year":"2023","journal-title":"Orphanet Journal of Rare Diseases"},{"key":"2024080900100562000_2023.03.19.23287458v4.22","unstructured":"Guney, E. Reproducible drug repurposing: When similarity does not suffice. In PSB (2017)."},{"key":"2024080900100562000_2023.03.19.23287458v4.23","doi-asserted-by":"crossref","first-page":"D1160","DOI":"10.1093\/nar\/gkaa997","article-title":"DrugCentral 2021 supports drug discovery and repositioning","volume":"49","year":"2021","journal-title":"Nucleic Acids Research"},{"key":"2024080900100562000_2023.03.19.23287458v4.24","unstructured":"Schlichtkrull, M. S. , De Cao, N. & Titov, I. Interpreting graph neural networks for NLP with differentiable edge masking. ICLR (2021)."},{"key":"2024080900100562000_2023.03.19.23287458v4.25","doi-asserted-by":"crossref","first-page":"1266","DOI":"10.1109\/TVCG.2022.3209435","article-title":"Extending the nested model for user-centric xai: A design study on gnn-based drug repurposing","volume":"29","year":"2023","journal-title":"IEEE Transactions on Visualization and Computer Graphics"},{"key":"2024080900100562000_2023.03.19.23287458v4.26","doi-asserted-by":"publisher","DOI":"10.1371\/journal.pone.0076339"},{"key":"2024080900100562000_2023.03.19.23287458v4.27","doi-asserted-by":"crossref","unstructured":"Schlichtkrull, M. et al. Modeling relational data with graph convolutional networks. In ESWC (2018).","DOI":"10.1007\/978-3-319-93417-4_38"},{"key":"2024080900100562000_2023.03.19.23287458v4.28","doi-asserted-by":"crossref","unstructured":"Hu, Z. , Dong, Y. , Wang, K. & Sun, Y. Heterogeneous graph transformer. WWW (2020).","DOI":"10.1145\/3366423.3380027"},{"key":"2024080900100562000_2023.03.19.23287458v4.29","doi-asserted-by":"crossref","unstructured":"Wang, X. et al. Heterogeneous graph attention network. WWW (2019).","DOI":"10.1145\/3308558.3313562"},{"key":"2024080900100562000_2023.03.19.23287458v4.30","doi-asserted-by":"crossref","unstructured":"Lee, J. et al. BioBERT: a pre-trained biomedical language representation model for biomedical text mining. Bioinformatics (2019).","DOI":"10.1093\/bioinformatics\/btz682"},{"key":"2024080900100562000_2023.03.19.23287458v4.31","doi-asserted-by":"publisher","DOI":"10.1038\/s41587-020-0564-6"},{"key":"2024080900100562000_2023.03.19.23287458v4.32","doi-asserted-by":"crossref","unstructured":"Bickel, S. , Br\u00fcckner, M. & Scheffer, T. Discriminative learning under covariate shift. Journal of Machine Learning Research 10 (2009).","DOI":"10.7551\/mitpress\/7921.003.0014"},{"key":"2024080900100562000_2023.03.19.23287458v4.33","doi-asserted-by":"crossref","unstructured":"Niven, T. & Kao, H.-Y. Probing neural network comprehension of natural language arguments. ACL (2019).","DOI":"10.18653\/v1\/P19-1459"},{"key":"2024080900100562000_2023.03.19.23287458v4.34","doi-asserted-by":"crossref","first-page":"665","DOI":"10.1038\/s42256-020-00257-z","article-title":"Shortcut learning in deep neural networks","volume":"2","year":"2020","journal-title":"Nature Machine Intelligence"},{"key":"2024080900100562000_2023.03.19.23287458v4.35","doi-asserted-by":"crossref","unstructured":"Agarwal, C. , Queen, O. , Lakkaraju, H. & Zitnik, M. Evaluating explainability for graph neural networks. Scientific Data 10 (2023).","DOI":"10.1038\/s41597-023-01974-x"},{"key":"2024080900100562000_2023.03.19.23287458v4.36","unstructured":"Ying, Z. , Bourgeois, D. , You, J. , Zitnik, M. & Leskovec, J. GNNExplainer: Generating explanations for graph neural networks. NeurIPS (2019)."},{"key":"2024080900100562000_2023.03.19.23287458v4.37","unstructured":"Sundararajan, M. , Taly, A. & Yan, Q. Axiomatic attribution for deep networks. In ICML (2017)."},{"key":"2024080900100562000_2023.03.19.23287458v4.38","doi-asserted-by":"crossref","unstructured":"Wang, J. et al. Empower post-hoc graph explanations with information bottleneck: A pretraining and fine-tuning perspective. In KDD (2023).","DOI":"10.1145\/3580305.3599330"},{"key":"2024080900100562000_2023.03.19.23287458v4.39","doi-asserted-by":"crossref","unstructured":"Tukey, J. W. Comparing individual means in the analysis of variance. Biometrics 99\u2013114 (1949).","DOI":"10.2307\/3001913"},{"key":"2024080900100562000_2023.03.19.23287458v4.40","doi-asserted-by":"crossref","first-page":"1130","DOI":"10.1001\/jamaneurol.2017.1133","article-title":"Zolpidem for the treatment of neurologic disorders: a systematic review","volume":"74","year":"2017","journal-title":"JAMA Neurology"},{"key":"2024080900100562000_2023.03.19.23287458v4.41","doi-asserted-by":"crossref","first-page":"1127542","DOI":"10.3389\/fnins.2023.1127542","article-title":"Case report: Zolpidem\u2019s paradoxical restorative action: A case report of functional brain imaging","volume":"17","year":"2023","journal-title":"Frontiers in Neuroscience"},{"key":"2024080900100562000_2023.03.19.23287458v4.42","doi-asserted-by":"crossref","unstructured":"Sripad, P. et al. Effect of zolpidem in the aftermath of traumatic brain injury: an MEG study. Case Reports in Neurological Medicine 2020 (2020).","DOI":"10.1155\/2020\/8597062"},{"key":"2024080900100562000_2023.03.19.23287458v4.43","doi-asserted-by":"publisher","DOI":"10.1093\/nar\/gkz972"},{"key":"2024080900100562000_2023.03.19.23287458v4.44","doi-asserted-by":"crossref","first-page":"367","DOI":"10.4103\/ijmr.IJMR_709_20","article-title":"ALDH1 & CD133 in invasive cervical carcinoma & their association with the outcome of chemoradiation therapy","volume":"154","year":"2021","journal-title":"The Indian Journal of Medical Research"},{"key":"2024080900100562000_2023.03.19.23287458v4.45","doi-asserted-by":"publisher","DOI":"10.1093\/NAR\/GKAA840"},{"key":"2024080900100562000_2023.03.19.23287458v4.46","doi-asserted-by":"crossref","first-page":"4779","DOI":"10.1111\/jcmm.14366","article-title":"Rosiglitazone treatment restores renal responsiveness to atrial natriuretic peptide in rats with congestive heart failure","volume":"23","year":"2019","journal-title":"Journal of Cellular and Molecular Medicine"},{"key":"2024080900100562000_2023.03.19.23287458v4.47","doi-asserted-by":"publisher","DOI":"10.1152\/ajprenal.00418.2006"},{"key":"2024080900100562000_2023.03.19.23287458v4.48","doi-asserted-by":"publisher","DOI":"10.1093\/nar\/gkx1037"},{"key":"2024080900100562000_2023.03.19.23287458v4.49","doi-asserted-by":"crossref","first-page":"1552","DOI":"10.4254\/wjh.v13.i11.1552","article-title":"Chelation therapy in liver diseases of childhood: Current status and response","volume":"13","year":"2021","journal-title":"World Journal of Hepatology"},{"key":"2024080900100562000_2023.03.19.23287458v4.50","unstructured":"Alsentzer, E. et al. Deep learning for diagnosing patients with rare genetic diseases. medRxiv 2022\u201312 (2022)."},{"key":"2024080900100562000_2023.03.19.23287458v4.51","unstructured":"Zhang, A. , Xing, L. , Zou, J. & Wu, J. C. Shifting machine learning for healthcare from development to deployment and from models to data. Nature Biomedical Engineering 1\u201316 (2022)."},{"key":"2024080900100562000_2023.03.19.23287458v4.52","doi-asserted-by":"crossref","unstructured":"Duffy, A. et al. Development of a human genetics-guided priority score for 19,365 genes and 399 drug indications. Nature Genetics 1\u20139 (2024).","DOI":"10.1038\/s41588-023-01609-2"},{"key":"2024080900100562000_2023.03.19.23287458v4.53","unstructured":"Cheng, J. , Dasoulas, G. , He, H. , Agarwal, C. & Zitnik, M. GNNDelete: a general strategy for unlearning in graph neural networks. ICLR (2023)"},{"key":"2024080900100562000_2023.03.19.23287458v4.54","unstructured":"Huang, K. , Jin, Y. , Candes, E. & Leskovec, J. Uncertainty quantification over graph with conformalized graph neural networks. NeurIPS (2024)."},{"key":"2024080900100562000_2023.03.19.23287458v4.55","doi-asserted-by":"crossref","unstructured":"Cai, C. J. et al. Human-centered tools for coping with imperfect algorithms during medical decision-making. In CHI (2019).","DOI":"10.1145\/3290605.3300234"},{"key":"2024080900100562000_2023.03.19.23287458v4.56","first-page":"34","article-title":"How to specify the participant group size for usability studies: a practitioner\u2019s guide","volume":"5","year":"2009","journal-title":"Journal of Usability Studies"},{"key":"2024080900100562000_2023.03.19.23287458v4.57","unstructured":"Gilmer, J. , Schoenholz, S. S. , Riley, P. F. , Vinyals, O. & Dahl, G. E. Neural message passing for quantum chemistry. In ICML (2017)."},{"key":"2024080900100562000_2023.03.19.23287458v4.58","unstructured":"Yang, B. , Yih, W.-t. , He, X. , Gao, J. & Deng, L. Embedding entities and relations for learning and inference in knowledge bases. ICLR (2015)."},{"key":"2024080900100562000_2023.03.19.23287458v4.59","doi-asserted-by":"crossref","unstructured":"Lin, Y. , Liu, Z. , Sun, M. , Liu, Y. & Zhu, X. Learning entity and relation embeddings for knowledge graph completion. In AAAI (2015).","DOI":"10.1609\/aaai.v29i1.9491"},{"key":"2024080900100562000_2023.03.19.23287458v4.60","doi-asserted-by":"publisher","DOI":"10.7326\/0003-4819-153-9-201011020-00010"}],"container-title":[],"original-title":[],"link":[{"URL":"https:\/\/syndication.highwire.org\/content\/doi\/10.1101\/2023.03.19.23287458","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,1,15]],"date-time":"2026-01-15T19:30:15Z","timestamp":1768505415000},"score":1,"resource":{"primary":{"URL":"http:\/\/medrxiv.org\/lookup\/doi\/10.1101\/2023.03.19.23287458"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2023,3,20]]},"references-count":60,"URL":"https:\/\/doi.org\/10.1101\/2023.03.19.23287458","relation":{"is-preprint-of":[{"id-type":"doi","id":"10.1038\/s41591-024-03233-x","asserted-by":"subject"}]},"subject":[],"published":{"date-parts":[[2023,3,20]]},"subtype":"preprint"}}