{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2025,9,24]],"date-time":"2025-09-24T09:08:21Z","timestamp":1758704901879,"version":"3.37.3"},"reference-count":71,"publisher":"Oxford University Press (OUP)","issue":"Supplement_1","license":[{"start":{"date-parts":[[2023,6,30]],"date-time":"2023-06-30T00:00:00Z","timestamp":1688083200000},"content-version":"vor","delay-in-days":29,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"DOI":"10.13039\/501100001659","name":"Deutsche Forschungsgemeinschaft","doi-asserted-by":"publisher","id":[{"id":"10.13039\/501100001659","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100001659","name":"German Research Foundation","doi-asserted-by":"publisher","award":["390727645"],"award-info":[{"award-number":["390727645"]}],"id":[{"id":"10.13039\/501100001659","id-type":"DOI","asserted-by":"publisher"}]},{"name":"German Federal Ministry of Education and Research"},{"name":"Training Center Machine Learning, T\u00fcbingen","award":["01-S17054"],"award-info":[{"award-number":["01-S17054"]}]},{"name":"German Federal Ministry of Education and Research"},{"name":"T\u00fcbingen AI Center","award":["01IS18039A"],"award-info":[{"award-number":["01IS18039A"]}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2023,6,30]]},"abstract":"<jats:title>Abstract<\/jats:title>\n               <jats:sec>\n                  <jats:title>Motivation<\/jats:title>\n                  <jats:p>The size of available omics datasets is steadily increasing with technological advancement in recent years. While this increase in sample size can be used to improve the performance of relevant prediction tasks in healthcare, models that are optimized for large datasets usually operate as black boxes. In high-stakes scenarios, like healthcare, using a black-box model poses safety and security issues. Without an explanation about molecular factors and phenotypes that affected the prediction, healthcare providers are left with no choice but to blindly trust the models. We propose a new type of artificial neural network, named Convolutional Omics Kernel Network (COmic). By combining convolutional kernel networks with pathway-induced kernels, our method enables robust and interpretable end-to-end learning on omics datasets ranging in size from a few hundred to several hundreds of thousands of samples. Furthermore, COmic can be easily adapted to utilize multiomics data.<\/jats:p>\n               <\/jats:sec>\n               <jats:sec>\n                  <jats:title>Results<\/jats:title>\n                  <jats:p>We evaluated the performance capabilities of COmic on six different breast cancer cohorts. Additionally, we trained COmic models on multiomics data using the METABRIC cohort. Our models performed either better or similar to competitors on both tasks. We show how the use of pathway-induced Laplacian kernels opens the black-box nature of neural networks and results in intrinsically interpretable models that eliminate the need for post hoc explanation models.<\/jats:p>\n               <\/jats:sec>\n               <jats:sec>\n                  <jats:title>Availability and implementation<\/jats:title>\n                  <jats:p>Datasets, labels, and pathway-induced graph Laplacians used for the single-omics tasks can be downloaded at https:\/\/ibm.ent.box.com\/s\/ac2ilhyn7xjj27r0xiwtom4crccuobst\/folder\/48027287036. While datasets and graph Laplacians for the METABRIC cohort can be downloaded from the above mentioned repository, the labels have to be downloaded from cBioPortal at https:\/\/www.cbioportal.org\/study\/clinicalData?id=brca\\_metabric. COmic source code as well as all scripts necessary to reproduce the experiments and analysis are publicly available at https:\/\/github.com\/jditz\/comics.<\/jats:p>\n               <\/jats:sec>","DOI":"10.1093\/bioinformatics\/btad204","type":"journal-article","created":{"date-parts":[[2023,6,30]],"date-time":"2023-06-30T08:16:14Z","timestamp":1688112974000},"page":"i76-i85","source":"Crossref","is-referenced-by-count":3,"title":["COmic: convolutional kernel networks for interpretable end-to-end learning on (multi-)omics data"],"prefix":"10.1093","volume":"39","author":[{"given":"Jonas C","family":"Ditz","sequence":"first","affiliation":[{"name":"Methods in Medical Informatics, Department of Computer Science, University of T\u00fcbingen , T\u00fcbingen 72076, Germany"}]},{"given":"Bernhard","family":"Reuter","sequence":"additional","affiliation":[{"name":"Methods in Medical Informatics, Department of Computer Science, University of T\u00fcbingen , T\u00fcbingen 72076, Germany"}]},{"given":"Nico","family":"Pfeifer","sequence":"additional","affiliation":[{"name":"Methods in Medical Informatics, Department of Computer Science, University of T\u00fcbingen , T\u00fcbingen 72076, Germany"}]}],"member":"286","published-online":{"date-parts":[[2023,6,30]]},"reference":[{"key":"2023063008151604300_btad204-B1","doi-asserted-by":"crossref","first-page":"831","DOI":"10.1038\/nbt.3300","article-title":"Predicting the sequence specificities of DNA-and RNA-binding proteins by deep learning","volume":"33","author":"Alipanahi","year":"2015","journal-title":"Nat Biotechnol"},{"key":"2023063008151604300_btad204-B2","doi-asserted-by":"crossref","first-page":"769","DOI":"10.1080\/17460441.2019.1621284","article-title":"Using artificial intelligence methods to speed up drug discovery","volume":"14","author":"\u00c1lvarez-Machancoses","year":"2019","journal-title":"Expert Opin Drug Disc"},{"key":"2023063008151604300_btad204-B3","doi-asserted-by":"crossref","first-page":"259","DOI":"10.2214\/AJR.18.20391","article-title":"A review of the role of augmented intelligence in breast imaging: from automated breast density assessment to risk stratification","volume":"212","author":"Arieno","year":"2019","journal-title":"Am J Roentgenol"},{"first-page":"1729","year":"2011","author":"Bo","key":"2023063008151604300_btad204-B4"},{"author":"Bordt","key":"2023063008151604300_btad204-B5"},{"key":"2023063008151604300_btad204-B6","doi-asserted-by":"crossref","first-page":"3294","DOI":"10.1093\/bioinformatics\/btz094","article-title":"Biological sequence modeling with convolutional kernel networks","volume":"35","author":"Chen","year":"2019","journal-title":"Bioinformatics"},{"key":"2023063008151604300_btad204-B7","first-page":"37:1576","article-title":"Convolutional kernel networks for graph-structured data","author":"Chen","year":"2020","journal-title":"Int Conf Mach Learn"},{"key":"2023063008151604300_btad204-B8","first-page":"32:13431","article-title":"Recurrent kernel networks","author":"Chen","year":"2019","journal-title":"Adv Neural Inf Process Syst"},{"key":"2023063008151604300_btad204-B9","doi-asserted-by":"crossref","first-page":"S1","DOI":"10.1186\/1752-0509-5-S3-S1","article-title":"Identifying cancer biomarkers by network-constrained support vector machines","volume":"5","author":"Chen","year":"2011","journal-title":"BMC Syst Biol"},{"key":"2023063008151604300_btad204-B10","first-page":"342","article-title":"Kernel methods for deep learning","volume":"22","author":"Cho","year":"2009","journal-title":"Adv Neural Inf Process Syst"},{"first-page":"9268","year":"2019","author":"Cui","key":"2023063008151604300_btad204-B11"},{"key":"2023063008151604300_btad204-B12","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1186\/1471-2105-13-69","article-title":"Prognostic gene signatures for patient stratification in breast cancer-accuracy, stability and interpretability of gene selection approaches using prior knowledge on protein-protein interactions","volume":"13","author":"Cun","year":"2012","journal-title":"BMC Bioinformatics"},{"key":"2023063008151604300_btad204-B13","doi-asserted-by":"crossref","first-page":"346","DOI":"10.1038\/nature10983","article-title":"The genomic and transcriptomic architecture of 2,000 breast tumours reveals novel subgroups","volume":"486","author":"Curtis","year":"2012","journal-title":"Nature"},{"first-page":"933","year":"2017","author":"Dauphin","key":"2023063008151604300_btad204-B14"},{"key":"2023063008151604300_btad204-B15","doi-asserted-by":"crossref","first-page":"1342","DOI":"10.1038\/s41591-018-0107-6","article-title":"Clinically applicable deep learning for diagnosis and referral in retinal disease","volume":"24","author":"De Fauw","year":"2018","journal-title":"Nat Med"},{"key":"2023063008151604300_btad204-B16","doi-asserted-by":"crossref","first-page":"3207","DOI":"10.1158\/1078-0432.CCR-06-2765","article-title":"Strong time dependence of the 76-gene prognostic signature for node-negative breast cancer patients in the transbig multicenter independent validation series","volume":"13","author":"Desmedt","year":"2007","journal-title":"Clin Cancer Res"},{"key":"2023063008151604300_btad204-B17","doi-asserted-by":"crossref","first-page":"31","DOI":"10.1016\/S0004-3702(96)00034-3","article-title":"Solving the multiple instance problem with axis-parallel rectangles","volume":"89","author":"Dietterich","year":"1997","journal-title":"Artif Intell"},{"year":"2021","author":"Ditz","key":"2023063008151604300_btad204-B18"},{"key":"2023063008151604300_btad204-B19","first-page":"15:1441","article-title":"Mismatch string kernels for SVM protein classification","author":"Eskin","year":"2003","journal-title":"Adv Neural Inf Process Syst"},{"key":"2023063008151604300_btad204-B20","doi-asserted-by":"crossref","first-page":"169","DOI":"10.1007\/s12015-007-0023-5","article-title":"Mammary stem cells and breast cancer\u2014role of notch signalling","volume":"3","author":"Farnie","year":"2007","journal-title":"Stem Cell Rev"},{"key":"2023063008151604300_btad204-B21","doi-asserted-by":"crossref","first-page":"629","DOI":"10.1016\/j.patcog.2016.07.016","article-title":"Bacterial colony counting with convolutional neural networks in digital microbiology imaging","volume":"61","author":"Ferrari","year":"2017","journal-title":"Pattern Recognit"},{"key":"2023063008151604300_btad204-B22","doi-asserted-by":"crossref","first-page":"S55","DOI":"10.1038\/d41586-018-05267-x","article-title":"How artificial intelligence is changing drug discovery","volume":"557","author":"Fleming","year":"2018","journal-title":"Nature"},{"key":"2023063008151604300_btad204-B23","doi-asserted-by":"crossref","first-page":"S19","DOI":"10.1186\/1471-2105-10-S11-S19","article-title":"Graph ranking for exploratory gene data analysis","volume":"10","author":"Gao","year":"2009","journal-title":"BMC Bioinformatics"},{"key":"2023063008151604300_btad204-B24","doi-asserted-by":"crossref","first-page":"58","DOI":"10.1186\/1471-2105-6-58","article-title":"Towards precise classification of cancers based on robust gene functional expression profiles","volume":"6","author":"Guo","year":"2005","journal-title":"BMC Bioinformatics"},{"key":"2023063008151604300_btad204-B25","doi-asserted-by":"crossref","first-page":"389","DOI":"10.1023\/A:1012487302797","article-title":"Gene selection for cancer classification using support vector machines","volume":"46","author":"Guyon","year":"2002","journal-title":"Mach Learn"},{"key":"2023063008151604300_btad204-B26","doi-asserted-by":"crossref","first-page":"e1001413","DOI":"10.1371\/journal.pmed.1001413","article-title":"Big data opportunities for global infectious disease surveillance","volume":"10","author":"Hay","year":"2013","journal-title":"PLoS Med"},{"first-page":"2127","year":"2018","author":"Ilse","key":"2023063008151604300_btad204-B27"},{"key":"2023063008151604300_btad204-B28","doi-asserted-by":"crossref","first-page":"10292","DOI":"10.1158\/0008-5472.CAN-05-4414","article-title":"Genetic reclassification of histologic grade delineates new clinical subtypes of breast cancer","volume":"66","author":"Ivshina","year":"2006","journal-title":"Cancer Res"},{"key":"2023063008151604300_btad204-B29","doi-asserted-by":"crossref","first-page":"134","DOI":"10.1158\/2643-3230.BCD-20-0007","article-title":"Hedgehog pathway inhibitors: a new therapeutic class for the treatment of acute myeloid leukemia hedgehog pathway inhibitors for acute myeloid leukemia","volume":"1","author":"Jamieson","year":"2020","journal-title":"Blood Cancer Discov"},{"key":"2023063008151604300_btad204-B30","doi-asserted-by":"crossref","first-page":"990","DOI":"10.1101\/gr.200535.115","article-title":"Basset: learning the regulatory code of the accessible genome with deep convolutional neural networks","volume":"26","author":"Kelley","year":"2016","journal-title":"Genome Res"},{"year":"2014","author":"Kingma","key":"2023063008151604300_btad204-B31"},{"key":"2023063008151604300_btad204-B32","doi-asserted-by":"crossref","first-page":"92974","DOI":"10.1109\/ACCESS.2021.3093456","article-title":"Attention meets perturbations: robust and interpretable attention with adversarial training","volume":"9","author":"Kitada","year":"2021","journal-title":"IEEE Access"},{"key":"2023063008151604300_btad204-B33","doi-asserted-by":"crossref","first-page":"e1000217","DOI":"10.1371\/journal.pcbi.1000217","article-title":"Inferring pathway activity toward precise disease classification","volume":"4","author":"Lee","year":"2008","journal-title":"PLoS Comput Biol"},{"key":"2023063008151604300_btad204-B34","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1145\/2948068","article-title":"The convergence behavior of naive bayes on large sparse datasets","volume":"11","author":"Li","year":"2016","journal-title":"ACM Trans Knowl Discov Data"},{"key":"2023063008151604300_btad204-B35","doi-asserted-by":"crossref","first-page":"31","DOI":"10.1145\/3236386.3241340","article-title":"The mythos of model interpretability: in machine learning, the concept of interpretability is both important and slippery","volume":"16","author":"Lipton","year":"2018","journal-title":"Queue"},{"first-page":"253","year":"2012","author":"Liu","key":"2023063008151604300_btad204-B36"},{"first-page":"4768","year":"2017","author":"Lundberg","key":"2023063008151604300_btad204-B37"},{"key":"2023063008151604300_btad204-B38","first-page":"29:1399","article-title":"End-to-end kernel learning with supervised convolutional kernel networks","author":"Mairal","year":"2016","journal-title":"Adv Neural Inf Process Syst"},{"key":"2023063008151604300_btad204-B39","first-page":"27:2627","article-title":"Convolutional kernel networks","author":"Mairal","year":"2014","journal-title":"Adv Neural Inf Process Syst"},{"key":"2023063008151604300_btad204-B40","doi-asserted-by":"crossref","first-page":"8","DOI":"10.1038\/s41540-019-0086-3","article-title":"Pimkl: pathway-induced multiple kernel learning","volume":"5","author":"Manica","year":"2019","journal-title":"NPJ Syst Biol Appl"},{"key":"2023063008151604300_btad204-B41","first-page":"10","article-title":"A framework for multiple-instance learning","author":"Maron","year":"1997","journal-title":"Adv Neural Inf Process Syst"},{"key":"2023063008151604300_btad204-B42","doi-asserted-by":"crossref","first-page":"169","DOI":"10.1186\/1471-2105-5-169","article-title":"Oligo kernels for datamining on biological sequences: a case study on prokaryotic translation initiation sites","volume":"5","author":"Meinicke","year":"2004","journal-title":"BMC Bioinformatics"},{"key":"2023063008151604300_btad204-B43","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1038\/srep26094","article-title":"Deep patient: an unsupervised representation to predict the future of patients from the electronic health records","volume":"6","author":"Miotto","year":"2016","journal-title":"Sci Rep"},{"volume-title":"Interpretable Machine Learning","year":"2020","author":"Molnar","key":"2023063008151604300_btad204-B44"},{"first-page":"193","year":"2019","author":"Montavon","key":"2023063008151604300_btad204-B45"},{"key":"2023063008151604300_btad204-B46","doi-asserted-by":"crossref","first-page":"211","DOI":"10.1016\/j.patcog.2016.11.008","article-title":"Explaining nonlinear classification decisions with deep taylor decomposition","volume":"65","author":"Montavon","year":"2017","journal-title":"Pattern Recognit"},{"key":"2023063008151604300_btad204-B47","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1038\/s41746-019-0103-3","article-title":"Predicting scheduled hospital attendance with artificial intelligence","volume":"2","author":"Nelson","year":"2019","journal-title":"NPJ Digit Med"},{"author":"Oquab","key":"2023063008151604300_btad204-B48","first-page":"685"},{"key":"2023063008151604300_btad204-B49","doi-asserted-by":"crossref","first-page":"1385","DOI":"10.1534\/g3.116.033654","article-title":"Accurate classification of protein subcellular localization from high-throughput microscopy images using deep learning","volume":"7","author":"P\u00e4rnamaa","year":"2017","journal-title":"G3 (Bethesda)"},{"key":"2023063008151604300_btad204-B50","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1186\/bcr1325","article-title":"Gene expression profiling spares early breast cancer patients from adjuvant therapy: derived and validated in two population-based cohorts","volume":"7","author":"Pawitan","year":"2005","journal-title":"Breast Cancer Res"},{"key":"2023063008151604300_btad204-B51","doi-asserted-by":"crossref","first-page":"R485","DOI":"10.1530\/ERC-16-0190","article-title":"Androgen receptor signaling pathways as a target for breast cancer treatment","volume":"23","author":"Pietri","year":"2016","journal-title":"Endocr Relat Cancer"},{"key":"2023063008151604300_btad204-B52","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1186\/1471-2105-8-35","article-title":"Classification of microarray data using gene networks","volume":"8","author":"Rapaport","year":"2007","journal-title":"BMC Bioinformatics"},{"first-page":"234","year":"2015","author":"Ronneberger","key":"2023063008151604300_btad204-B53"},{"key":"2023063008151604300_btad204-B54","doi-asserted-by":"crossref","first-page":"206","DOI":"10.1038\/s42256-019-0048-x","article-title":"Stop explaining black box machine learning models for high stakes decisions and use interpretable models instead","volume":"1","author":"Rudin","year":"2019","journal-title":"Nat Mach Intell"},{"key":"2023063008151604300_btad204-B55","doi-asserted-by":"crossref","first-page":"5405","DOI":"10.1158\/0008-5472.CAN-07-5206","article-title":"The humoral immune system has a key prognostic impact in node-negative breast cancer","volume":"68","author":"Schmidt","year":"2008","journal-title":"Cancer Res"},{"first-page":"3145","year":"2017","author":"Shrikumar","key":"2023063008151604300_btad204-B56"},{"first-page":"9046","year":"2020","author":"Sixt","key":"2023063008151604300_btad204-B57"},{"key":"2023063008151604300_btad204-B58","doi-asserted-by":"crossref","first-page":"262","DOI":"10.1093\/jnci\/djj052","article-title":"Gene expression profiling in breast cancer: understanding the molecular basis of histologic grade to improve prognosis","volume":"98","author":"Sotiriou","year":"2006","journal-title":"J Natl Cancer Inst"},{"year":"2014","author":"Springenberg","key":"2023063008151604300_btad204-B59"},{"key":"2023063008151604300_btad204-B60","doi-asserted-by":"crossref","first-page":"11974","DOI":"10.1109\/ACCESS.2021.3051315","article-title":"A survey of contrastive and counterfactual explanation generation methods for explainable artificial intelligence","volume":"9","author":"Stepin","year":"2021","journal-title":"IEEE Access"},{"key":"2023063008151604300_btad204-B61","doi-asserted-by":"crossref","first-page":"199","DOI":"10.1038\/nbt.1522","article-title":"Dynamic modularity in protein interaction networks predicts breast cancer outcome","volume":"27","author":"Taylor","year":"2009","journal-title":"Nat Biotechnol"},{"first-page":"1627","year":"2021","author":"Tran","key":"2023063008151604300_btad204-B62"},{"first-page":"1027","year":"2006","author":"Vassilvitskii","key":"2023063008151604300_btad204-B63"},{"key":"2023063008151604300_btad204-B64","doi-asserted-by":"crossref","first-page":"395","DOI":"10.1007\/s11222-007-9033-z","article-title":"A tutorial on spectral clustering","volume":"17","author":"Von Luxburg","year":"2007","journal-title":"Stat Comput"},{"key":"2023063008151604300_btad204-B65","doi-asserted-by":"crossref","first-page":"15","DOI":"10.1016\/j.patcog.2017.08.026","article-title":"Revisiting multiple instance neural networks","volume":"74","author":"Wang","year":"2018","journal-title":"Pattern Recognit"},{"key":"2023063008151604300_btad204-B66","doi-asserted-by":"crossref","first-page":"671","DOI":"10.1016\/S0140-6736(05)17947-1","article-title":"Gene-expression profiles to predict distant metastasis of lymph-node-negative primary breast cancer","volume":"365","author":"Wang","year":"2005","journal-title":"Lancet"},{"first-page":"682","year":"2001","author":"Williams","key":"2023063008151604300_btad204-B67"},{"key":"2023063008151604300_btad204-B68","doi-asserted-by":"crossref","first-page":"629","DOI":"10.1177\/1947601910378691","article-title":"Myc and breast cancer","volume":"1","author":"Xu","year":"2010","journal-title":"Genes Cancer"},{"first-page":"1232","year":"2008","author":"Zhang","key":"2023063008151604300_btad204-B69"},{"key":"2023063008151604300_btad204-B70","doi-asserted-by":"crossref","first-page":"931","DOI":"10.1038\/nmeth.3547","article-title":"Predicting effects of noncoding variants with deep learning\u2013based sequence model","volume":"12","author":"Zhou","year":"2015","journal-title":"Nat Methods"},{"key":"2023063008151604300_btad204-B71","doi-asserted-by":"crossref","first-page":"S21","DOI":"10.1186\/1471-2105-10-S1-S21","article-title":"Network-based support vector machine for classification of microarray samples","volume":"10","author":"Zhu","year":"2009","journal-title":"BMC Bioinformatics"}],"container-title":["Bioinformatics"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/academic.oup.com\/bioinformatics\/article-pdf\/39\/Supplement_1\/i76\/50741537\/btad204.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/academic.oup.com\/bioinformatics\/article-pdf\/39\/Supplement_1\/i76\/50741537\/btad204.pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2023,6,30]],"date-time":"2023-06-30T08:18:14Z","timestamp":1688113094000},"score":1,"resource":{"primary":{"URL":"https:\/\/academic.oup.com\/bioinformatics\/article\/39\/Supplement_1\/i76\/7210453"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2023,6,1]]},"references-count":71,"journal-issue":{"issue":"Supplement_1","published-print":{"date-parts":[[2023,6,30]]}},"URL":"https:\/\/doi.org\/10.1093\/bioinformatics\/btad204","relation":{},"ISSN":["1367-4803","1367-4811"],"issn-type":[{"type":"print","value":"1367-4803"},{"type":"electronic","value":"1367-4811"}],"subject":[],"published-other":{"date-parts":[[2023,6,1]]},"published":{"date-parts":[[2023,6,1]]}}}