{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,6,27]],"date-time":"2026-06-27T05:42:36Z","timestamp":1782538956260,"version":"3.54.5"},"reference-count":67,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2023,8,2]],"date-time":"2023-08-02T00:00:00Z","timestamp":1690934400000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"DOI":"10.13039\/100000002","name":"National Institutes of Health","doi-asserted-by":"publisher","award":["R35GM147241-01"],"award-info":[{"award-number":["R35GM147241-01"]}],"id":[{"id":"10.13039\/100000002","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Bioinform."],"abstract":"<jats:p><jats:bold>Introduction:<\/jats:bold> Existing large-scale preclinical cancer drug response databases provide us with a great opportunity to identify and predict potentially effective drugs to combat cancers. Deep learning models built on these databases have been developed and applied to tackle the cancer drug-response prediction task. Their prediction has been demonstrated to significantly outperform traditional machine learning methods. However, due to the \u201cblack box\u201d characteristic, biologically faithful explanations are hardly derived from these deep learning models. Interpretable deep learning models that rely on visible neural networks (VNNs) have been proposed to provide biological justification for the predicted outcomes. However, their performance does not meet the expectation to be applied in clinical practice.<\/jats:p><jats:p><jats:bold>Methods:<\/jats:bold> In this paper, we develop an XMR model, an eXplainable Multimodal neural network for drug Response prediction. XMR is a new compact multimodal neural network consisting of two sub-networks: a visible neural network for learning genomic features and a graph neural network (GNN) for learning drugs\u2019 structural features. Both sub-networks are integrated into a multimodal fusion layer to model the drug response for the given gene mutations and the drug\u2019s molecular structures. Furthermore, a pruning approach is applied to provide better interpretations of the XMR model. We use five pathway hierarchies (cell cycle, DNA repair, diseases, signal transduction, and metabolism), which are obtained from the Reactome Pathway Database, as the architecture of VNN for our XMR model to predict drug responses of triple negative breast cancer.<\/jats:p><jats:p><jats:bold>Results:<\/jats:bold> We find that our model outperforms other state-of-the-art interpretable deep learning models in terms of predictive performance. In addition, our model can provide biological insights into explaining drug responses for triple-negative breast cancer.<\/jats:p><jats:p><jats:bold>Discussion:<\/jats:bold> Overall, combining both VNN and GNN in a multimodal fusion layer, XMR captures key genomic and molecular features and offers reasonable interpretability in biology, thereby better predicting drug responses in cancer patients. Our model would also benefit personalized cancer therapy in the future.<\/jats:p>","DOI":"10.3389\/fbinf.2023.1164482","type":"journal-article","created":{"date-parts":[[2023,8,2]],"date-time":"2023-08-02T12:14:10Z","timestamp":1690978450000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":12,"title":["XMR: an explainable multimodal neural network for drug response prediction"],"prefix":"10.3389","volume":"3","author":[{"given":"Zihao","family":"Wang","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yun","family":"Zhou","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yu","family":"Zhang","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yu K.","family":"Mo","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yijie","family":"Wang","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1965","published-online":{"date-parts":[[2023,8,2]]},"reference":[{"key":"B1","doi-asserted-by":"publisher","first-page":"93","DOI":"10.1186\/s13058-017-0878-6","article-title":"Selinexor (kpt-330) demonstrates anti-tumor efficacy in preclinical models of triple-negative breast cancer","volume":"19","author":"Arango","year":"2017","journal-title":"Breast cancer Res."},{"key":"B2","doi-asserted-by":"publisher","first-page":"603","DOI":"10.1038\/nature11003","article-title":"The cancer cell line encyclopedia enables predictive modelling of anticancer drug sensitivity","volume":"483","author":"Barretina","year":"2012","journal-title":"Nature"},{"key":"B3","doi-asserted-by":"publisher","first-page":"1151","DOI":"10.1016\/j.cell.2013.08.003","article-title":"An interactive resource to identify cancer genetic and lineage dependencies targeted by small molecules","volume":"154","author":"Basu","year":"2013","journal-title":"Cell"},{"key":"B4","doi-asserted-by":"publisher","first-page":"355","DOI":"10.1038\/nature12627","article-title":"Tumour heterogeneity in the clinic","volume":"501","author":"Bedard","year":"2013","journal-title":"Nature"},{"key":"B5","doi-asserted-by":"publisher","first-page":"459","DOI":"10.1007\/s10107-013-0701-9","article-title":"Proximal alternating linearized minimization for nonconvex and nonsmooth problems","volume":"146","author":"Bolte","year":"2014","journal-title":"Math. Program."},{"key":"B6","doi-asserted-by":"publisher","first-page":"678","DOI":"10.1038\/nrm2009","article-title":"Notch signalling: A simple pathway becomes complex","volume":"7","author":"Bray","year":"2006","journal-title":"Nat. Rev. Mol. Cell Biol."},{"key":"B7","doi-asserted-by":"publisher","first-page":"34","DOI":"10.1186\/s12943-018-0797-x","article-title":"Receptor tyrosine kinases (rtks) in breast cancer: Signaling, therapeutic implications and challenges","volume":"17","author":"Butti","year":"2018","journal-title":"Mol. cancer"},{"key":"B8","doi-asserted-by":"publisher","first-page":"205","DOI":"10.21037\/atm-20-5989","article-title":"Identification of novel prognostic genes of triple-negative breast cancer using meta-analysis and weighted gene co-expressed network analysis","volume":"9","author":"Cao","year":"2021","journal-title":"Ann. Transl. Med."},{"key":"B9","doi-asserted-by":"publisher","first-page":"35","DOI":"10.3233\/bd-2010-0307","article-title":"Triple negative breast cancer cell lines: One tool in the search for better treatment of triple negative breast cancer","volume":"32","author":"Chavez","year":"2010","journal-title":"Breast Dis."},{"key":"B10","first-page":"104","article-title":"Uniter: Universal image-text representation learning","author":"Chen","year":"2020"},{"key":"B11","doi-asserted-by":"publisher","first-page":"675","DOI":"10.1158\/1535-7163.MCT-13-0416","article-title":"Xpo1 (crm1) inhibition represses stat3 activation to drive a survivin-dependent oncogenic switch in triple-negative breast cancer","volume":"13","author":"Cheng","year":"2014","journal-title":"Mol. cancer Ther."},{"key":"B12","first-page":"255","article-title":"Fast neighborhood subgraph pairwise distance kernel","volume-title":"Proceedings of the 26th international conference on machine learning","author":"Costa","year":"2010"},{"key":"B13","doi-asserted-by":"publisher","first-page":"1902","DOI":"10.1002\/ijc.32266","article-title":"A large collection of integrated genomically characterized patient-derived xenografts highlighting the heterogeneity of triple-negative breast cancer","volume":"145","author":"Coussy","year":"2019","journal-title":"Int. J. cancer"},{"key":"B14","doi-asserted-by":"publisher","first-page":"81","DOI":"10.1038\/nrclinonc.2017.166","article-title":"Tumour heterogeneity and resistance to cancer therapies","volume":"15","author":"Dagogo-Jack","year":"2018","journal-title":"Nat. Rev. Clin. Oncol."},{"key":"B15","doi-asserted-by":"publisher","first-page":"3131","DOI":"10.7150\/jca.18457","article-title":"Breast cancer cell line classification and its relevance with breast tumor subtyping","volume":"8","author":"Dai","year":"2017","journal-title":"J. Cancer"},{"key":"B16","doi-asserted-by":"publisher","article-title":"DepMap 22Q4 public","year":"2022","DOI":"10.6084\/m9.figshare.21637199.v2"},{"key":"B17","doi-asserted-by":"publisher","first-page":"e0121945","DOI":"10.1371\/journal.pone.0121945","article-title":"cocor: a comprehensive solution for the statistical comparison of correlations","volume":"10","author":"Diedenhofen","year":"2015","journal-title":"PLoS One"},{"key":"B18","doi-asserted-by":"publisher","first-page":"684316","DOI":"10.3389\/fonc.2021.684316","article-title":"Cardiac glycoside ouabain exerts anticancer activity via downregulation of stat3","volume":"11","author":"Du","year":"2021","journal-title":"Front. Oncol."},{"key":"B19","doi-asserted-by":"publisher","first-page":"68","DOI":"10.1145\/3359786","article-title":"Techniques for interpretable machine learning","volume":"63","author":"Du","year":"2020","journal-title":"Commun. ACM"},{"key":"B20","doi-asserted-by":"publisher","first-page":"348","DOI":"10.1038\/s41586-021-03922-4","article-title":"Biologically informed deep neural network for prostate cancer discovery","volume":"598","author":"Elmarakeby","year":"2021","journal-title":"Nature"},{"key":"B21","doi-asserted-by":"publisher","first-page":"D649","DOI":"10.1093\/nar\/gkx1132","article-title":"The reactome pathway knowledgebase","volume":"46","author":"Fabregat","year":"2018","journal-title":"Nucleic acids Res."},{"key":"B22","doi-asserted-by":"publisher","first-page":"160","DOI":"10.3389\/fcell.2020.00160","article-title":"Fanconi anemia pathway: Mechanisms of breast cancer predisposition development and potential therapeutic targets","volume":"8","author":"Fang","year":"2020","journal-title":"Front. Cell Dev. Biol."},{"key":"B23","doi-asserted-by":"publisher","first-page":"20","DOI":"10.1186\/s13073-019-0632-z","article-title":"Translating insights into tumor evolution to clinical practice: Promises and challenges","volume":"11","author":"Fittall","year":"2019","journal-title":"Genome Med."},{"key":"B24","doi-asserted-by":"publisher","first-page":"e0224420","DOI":"10.1371\/journal.pone.0224420","article-title":"Polo-like kinase 1 (plk1) inhibition synergizes with taxanes in triple negative breast cancer","volume":"14","author":"Giordano","year":"2019","journal-title":"PloS one"},{"key":"B25","doi-asserted-by":"publisher","first-page":"51","DOI":"10.1016\/j.cmet.2020.10.012","article-title":"Metabolic-pathway-based subtyping of triple-negative breast cancer reveals potential therapeutic targets","volume":"33","author":"Gong","year":"2021","journal-title":"Cell metab."},{"key":"B26","doi-asserted-by":"publisher","first-page":"2098","DOI":"10.1038\/cdd.2015.73","article-title":"Myeloid cell leukemia-1 is an important apoptotic survival factor in triple-negative breast cancer","volume":"22","author":"Goodwin","year":"2015","journal-title":"Cell death Differ."},{"key":"B27","doi-asserted-by":"publisher","first-page":"648","DOI":"10.1186\/s12885-021-08359-6","article-title":"Predicting breast cancer drug response using a multiple-layer cell line drug response network model","volume":"21","author":"Huang","year":"","journal-title":"BMC cancer"},{"key":"B28","doi-asserted-by":"publisher","first-page":"lqab097","DOI":"10.1093\/nargab\/lqab097","article-title":"ParsVNN: Parsimony visible neural networks for uncovering cancer-specific and drug-sensitive genes and pathways","volume":"3","author":"Huang","year":"","journal-title":"NAR Genomics Bioinforma."},{"key":"B29","doi-asserted-by":"publisher","first-page":"740","DOI":"10.1016\/j.cell.2016.06.017","article-title":"A landscape of pharmacogenomic interactions in cancer","volume":"166","author":"Iorio","year":"2016","journal-title":"Cell"},{"key":"B30","doi-asserted-by":"publisher","first-page":"1842","DOI":"10.3892\/ol.2020.11274","article-title":"Expression and clinical significance of mapk and egfr in triple-negative breast cancer","volume":"19","author":"Jiang","year":"2020","journal-title":"Oncol. Lett."},{"key":"B31","doi-asserted-by":"publisher","first-page":"728","DOI":"10.1038\/s41467-019-13825-8","article-title":"A deep learning system accurately classifies primary and metastatic cancers using passenger mutation patterns","volume":"11","author":"Jiao","year":"2020","journal-title":"Nat. Commun."},{"key":"B32","doi-asserted-by":"publisher","first-page":"D1373","DOI":"10.1093\/nar\/gkac956","article-title":"PubChem 2023 update","volume":"51","author":"Kim","year":"2022","journal-title":"Nucleic Acids Res."},{"key":"B33","first-page":"5583","article-title":"Vilt: Vision-and-language transformer without convolution or region supervision","author":"Kim","year":"2021"},{"key":"B34","doi-asserted-by":"publisher","first-page":"14937","DOI":"10.3390\/ijms232314937","article-title":"Genetic heterogeneity, tumor microenvironment and immunotherapy in triple-negative breast cancer","volume":"23","author":"Kudelova","year":"2022","journal-title":"Int. J. Mol. Sci."},{"key":"B35","doi-asserted-by":"publisher","first-page":"672","DOI":"10.1016\/j.ccell.2020.09.014","article-title":"Predicting drug response and synergy using a deep learning model of human cancer cells","volume":"38","author":"Kuenzi","year":"2020","journal-title":"Cancer Cell"},{"key":"B36","doi-asserted-by":"publisher","first-page":"e0223725","DOI":"10.1371\/journal.pone.0223725","article-title":"Defective base excision repair in the response to dna damaging agents in triple negative breast cancer","volume":"14","author":"Lee","year":"2019","journal-title":"PLoS One"},{"key":"B37","doi-asserted-by":"publisher","first-page":"2750","DOI":"10.1172\/JCI45014","article-title":"Identification of human triple-negative breast cancer subtypes and preclinical models for selection of targeted therapies","volume":"121","author":"Lehmann","year":"2011","journal-title":"J. Clin. investigation"},{"key":"B38","doi-asserted-by":"publisher","first-page":"6276","DOI":"10.1038\/s41467-021-26502-6","article-title":"Multi-omics analysis identifies therapeutic vulnerabilities in triple-negative breast cancer subtypes","volume":"12","author":"Lehmann","year":"2021","journal-title":"Nat. Commun."},{"key":"B39","first-page":"9694","article-title":"Align before fuse: Vision and language representation learning with momentum distillation","author":"Li","year":"2021"},{"key":"B40","doi-asserted-by":"publisher","first-page":"830124","DOI":"10.3389\/fonc.2022.830124","article-title":"Clinicopathologic and genomic features in triple-negative breast cancer between special and no-special morphologic pattern","volume":"12","author":"Li","year":"2022","journal-title":"Front. Oncol."},{"key":"B41","doi-asserted-by":"publisher","first-page":"134","DOI":"10.1186\/s13045-019-0818-2","article-title":"Emerging insights of tumor heterogeneity and drug resistance mechanisms in lung cancer targeted therapy","volume":"12","author":"Lim","year":"2019","journal-title":"J. Hematol. Oncol."},{"key":"B42","doi-asserted-by":"publisher","first-page":"861","DOI":"10.1038\/nrc2248","article-title":"The impact of a negligent g2\/m checkpoint on genomic instability and cancer induction","volume":"7","author":"L\u00f6brich","year":"2007","journal-title":"Nat. Rev. Cancer"},{"key":"B43","first-page":"13","article-title":"Vilbert: Pretraining task-agnostic visiolinguistic representations for vision-and-language tasks","author":"Lu","year":"2019"},{"key":"B44","doi-asserted-by":"publisher","first-page":"13","DOI":"10.3389\/fgene.2019.00013","article-title":"deepdriver: predicting cancer driver genes based on somatic mutations using deep convolutional neural networks","volume":"10","author":"Luo","year":"2019","journal-title":"Front. Genet."},{"key":"B45","doi-asserted-by":"publisher","first-page":"290","DOI":"10.1038\/nmeth.4627","article-title":"Using deep learning to model the hierarchical structure and function of a cell","volume":"15","author":"Ma","year":"2018","journal-title":"Nat. methods"},{"key":"B46","doi-asserted-by":"publisher","first-page":"624324","DOI":"10.3389\/fimmu.2021.624324","article-title":"Targeting slc1a5 and slc3a2\/slc7a5 as a potential strategy to strengthen anti-tumor immunity in the tumor microenvironment","volume":"12","author":"Nachef","year":"2021","journal-title":"Front. Immunol."},{"key":"B47","doi-asserted-by":"publisher","first-page":"957","DOI":"10.3390\/cells8090957","article-title":"Mechanisms of chemotherapy resistance in triple-negative breast cancer-how we can rise to the challenge","volume":"8","author":"Nedeljkovi\u0107","year":"2019","journal-title":"Cells"},{"key":"B48","doi-asserted-by":"publisher","first-page":"8","DOI":"10.1186\/s13045-020-01015-9","article-title":"Targeting oncogenic notch signaling with serca inhibitors","volume":"14","author":"Pagliaro","year":"2021","journal-title":"J. Hematol. Oncol."},{"key":"B49","doi-asserted-by":"publisher","first-page":"e0238262","DOI":"10.1371\/journal.pone.0238262","article-title":"Association between tumor mutation profile and clinical outcomes among hispanic latina women with triple-negative breast cancer","volume":"15","author":"Philipovskiy","year":"2020","journal-title":"PLoS One"},{"key":"B50","doi-asserted-by":"publisher","first-page":"15663","DOI":"10.1038\/s41598-022-19621-7","article-title":"Homoharringtonine demonstrates a cytotoxic effect against triple-negative breast cancer cell lines and acts synergistically with paclitaxel","volume":"12","author":"Plett","year":"2022","journal-title":"Sci. Rep."},{"key":"B51","doi-asserted-by":"publisher","first-page":"17899","DOI":"10.1038\/s41598-020-74882-4","article-title":"Pharmacologic profiling of patient-derived xenograft models of primary treatment-na\u00efve triple-negative breast cancer","volume":"10","author":"Powell","year":"2020","journal-title":"Sci. Rep."},{"key":"B52","first-page":"8748","article-title":"Learning transferable visual models from natural language supervision","author":"Radford","year":"2021"},{"key":"B53","doi-asserted-by":"publisher","first-page":"30","DOI":"10.21037\/acr-20-54","article-title":"Use of electrochemotherapy in a voluminous chest wall recurrence of triple-negative breast cancer: Case report","volume":"4","author":"Radica","year":"2020","journal-title":"AME case Rep."},{"key":"B54","doi-asserted-by":"publisher","first-page":"56864","DOI":"10.18632\/oncotarget.10870","article-title":"Inhibition of cyclin dependent kinase 9 by dinaciclib suppresses cyclin b1 expression and tumor growth in triple negative breast cancer","volume":"7","author":"Rajput","year":"2016","journal-title":"Oncotarget"},{"key":"B55","doi-asserted-by":"publisher","first-page":"161","DOI":"10.1007\/s00109-020-01874-2","article-title":"Clinical implications of intratumor heterogeneity: Challenges and opportunities","volume":"98","author":"Ram\u00f3n y Cajal","year":"2020","journal-title":"J. Mol. Med."},{"key":"B56","doi-asserted-by":"publisher","first-page":"1210","DOI":"10.1158\/2159-8290.cd-15-0235","article-title":"Harnessing connectivity in a large-scale small-molecule sensitivity dataset","volume":"5","author":"Seashore-Ludlow","year":"2015","journal-title":"Cancer Discov."},{"key":"B57","doi-asserted-by":"publisher","first-page":"428","DOI":"10.3389\/fonc.2020.00428","article-title":"Metabolic reprogramming in triple-negative breast cancer","volume":"10","author":"Sun","year":"2020","journal-title":"Front. Oncol."},{"key":"B58","doi-asserted-by":"publisher","first-page":"17256","DOI":"10.1038\/s41598-019-53989-3","article-title":"Identification of 12 cancer types through genome deep learning","volume":"9","author":"Sun","year":"2019","journal-title":"Sci. Rep."},{"key":"B59","doi-asserted-by":"publisher","first-page":"7682","DOI":"10.3390\/ijms22147682","article-title":"The role of pi3k\/akt and mapk signaling pathways in erythropoietin signalization","volume":"22","author":"T\u00f3thov\u00e1","year":"2021","journal-title":"Int. J. Mol. Sci."},{"key":"B60","doi-asserted-by":"publisher","first-page":"309","DOI":"10.1093\/bioinformatics\/bty535","article-title":"Compound\u2013protein interaction prediction with end-to-end learning of neural networks for graphs and sequences","volume":"35","author":"Tsubaki","year":"2019","journal-title":"Bioinformatics"},{"key":"B61","doi-asserted-by":"publisher","first-page":"6","DOI":"10.21037\/sci.2019.02.03","article-title":"The past, present, and future of crm1\/xpo1 inhibitors","volume":"6","author":"Wang","year":"2019","journal-title":"Stem Cell Investig."},{"key":"B62","doi-asserted-by":"publisher","first-page":"eaat8464","DOI":"10.1126\/science.aat8464","article-title":"Comprehensive functional genomic resource and integrative model for the human brain","volume":"362","author":"Wang","year":"2018","journal-title":"Science"},{"key":"B63","first-page":"1043","article-title":"Homoharringtonine, an approved anti-leukemia drug, suppresses triple negative breast cancer growth through a rapid reduction of anti-apoptotic protein abundance","volume":"9","author":"Yakhni","year":"2019","journal-title":"Am. J. cancer Res."},{"key":"B64","doi-asserted-by":"publisher","first-page":"D955","DOI":"10.1093\/nar\/gks1111","article-title":"Genomics of drug sensitivity in cancer (gdsc): A resource for therapeutic biomarker discovery in cancer cells","volume":"41","author":"Yang","year":"2012","journal-title":"Nucleic acids Res."},{"key":"B65","doi-asserted-by":"publisher","first-page":"476","DOI":"10.1186\/s12859-016-1334-9","article-title":"Deepgene: An advanced cancer type classifier based on deep learning and somatic point mutations","volume":"17","author":"Yuan","year":"2016","journal-title":"BMC Bioinforma."},{"key":"B66","doi-asserted-by":"publisher","first-page":"174366","DOI":"10.1016\/j.ejphar.2021.174366","article-title":"One therapeutic approach for triple-negative breast cancer: Checkpoint kinase 1 inhibitor azd7762 combination with neoadjuvant carboplatin","volume":"908","author":"Zhu","year":"2021","journal-title":"Eur. J. Pharmacol."},{"key":"B67","doi-asserted-by":"publisher","first-page":"188678","DOI":"10.1016\/j.bbcan.2022.188678","article-title":"Platinum-based systematic therapy in triple-negative breast cancer","volume":"1877","author":"Zhu","year":"2022","journal-title":"Rev. cancer"}],"container-title":["Frontiers in Bioinformatics"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fbinf.2023.1164482\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2023,8,2]],"date-time":"2023-08-02T12:14:20Z","timestamp":1690978460000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fbinf.2023.1164482\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2023,8,2]]},"references-count":67,"alternative-id":["10.3389\/fbinf.2023.1164482"],"URL":"https:\/\/doi.org\/10.3389\/fbinf.2023.1164482","relation":{},"ISSN":["2673-7647"],"issn-type":[{"value":"2673-7647","type":"electronic"}],"subject":[],"published":{"date-parts":[[2023,8,2]]},"article-number":"1164482"}}