{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,7,24]],"date-time":"2026-07-24T04:05:51Z","timestamp":1784865951356,"version":"3.55.0"},"update-to":[{"DOI":"10.1371\/journal.pcbi.1011240","type":"new_version","label":"New version","source":"publisher","updated":{"date-parts":[[2023,7,13]],"date-time":"2023-07-13T00:00:00Z","timestamp":1689206400000}}],"reference-count":59,"publisher":"Public Library of Science (PLoS)","issue":"6","license":[{"start":{"date-parts":[[2023,6,30]],"date-time":"2023-06-30T00:00:00Z","timestamp":1688083200000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"DOI":"10.13039\/501100009708","name":"Novo Nordisk Fonden","doi-asserted-by":"publisher","award":["0069071"],"award-info":[{"award-number":["0069071"]}],"id":[{"id":"10.13039\/501100009708","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100009708","name":"Novo Nordisk Fonden","doi-asserted-by":"publisher","award":["0069071"],"award-info":[{"award-number":["0069071"]}],"id":[{"id":"10.13039\/501100009708","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100009318","name":"Helmholtz Association","doi-asserted-by":"publisher","id":[{"id":"10.13039\/501100009318","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["www.ploscompbiol.org"],"crossmark-restriction":false},"short-container-title":["PLoS Comput Biol"],"abstract":"<jats:p>Supervised learning, such as regression and classification, is an essential tool for analyzing modern high-throughput sequencing data, for example in microbiome research. However, due to the compositionality and sparsity, existing techniques are often inadequate. Either they rely on extensions of the linear log-contrast model (which adjust for compositionality but cannot account for complex signals or sparsity) or they are based on black-box machine learning methods (which may capture useful signals, but lack interpretability due to the compositionality). We propose<jats:monospace specific-use=\"no-wrap\">KernelBiome<\/jats:monospace>, a kernel-based nonparametric regression and classification framework for compositional data. It is tailored to sparse compositional data and is able to incorporate prior knowledge, such as phylogenetic structure.<jats:monospace specific-use=\"no-wrap\">KernelBiome<\/jats:monospace>captures complex signals, including in the zero-structure, while automatically adapting model complexity. We demonstrate on par or improved predictive performance compared with state-of-the-art machine learning methods on 33 publicly available microbiome datasets. Additionally, our framework provides two key advantages: (i) We propose two novel quantities to interpret contributions of individual components and prove that they consistently estimate average perturbation effects of the conditional mean, extending the interpretability of linear log-contrast coefficients to nonparametric models. (ii) We show that the connection between kernels and distances aids interpretability and provides a data-driven embedding that can augment further analysis.<jats:monospace specific-use=\"no-wrap\">KernelBiome<\/jats:monospace>is available as an open-source Python package on PyPI and at<jats:ext-link xmlns:xlink=\"http:\/\/www.w3.org\/1999\/xlink\" ext-link-type=\"uri\" xlink:href=\"https:\/\/github.com\/shimenghuang\/KernelBiome\" xlink:type=\"simple\">https:\/\/github.com\/shimenghuang\/KernelBiome<\/jats:ext-link>.<\/jats:p>","DOI":"10.1371\/journal.pcbi.1011240","type":"journal-article","created":{"date-parts":[[2023,6,30]],"date-time":"2023-06-30T17:34:35Z","timestamp":1688146475000},"page":"e1011240","update-policy":"https:\/\/doi.org\/10.1371\/journal.pcbi.corrections_policy","source":"Crossref","is-referenced-by-count":14,"title":["Supervised learning and model analysis with compositional data"],"prefix":"10.1371","volume":"19","author":[{"ORCID":"https:\/\/orcid.org\/0000-0001-6919-821X","authenticated-orcid":true,"given":"Shimeng","family":"Huang","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0008-6358-699X","authenticated-orcid":true,"given":"Elisabeth","family":"Ailer","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-8718-4305","authenticated-orcid":true,"given":"Niki","family":"Kilbertus","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-6203-9777","authenticated-orcid":true,"given":"Niklas","family":"Pfister","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"340","published-online":{"date-parts":[[2023,6,30]]},"reference":[{"key":"pcbi.1011240.ref001","volume-title":"GSL Special Publications","author":"A Buccianti","year":"2006"},{"issue":"2","key":"pcbi.1011240.ref002","doi-asserted-by":"crossref","first-page":"130","DOI":"10.1021\/co5001458","article-title":"Statistical analysis and interpolation of compositional data in materials science","volume":"17","author":"MZ Pesenson","year":"2015","journal-title":"ACS combinatorial science"},{"issue":"3","key":"pcbi.1011240.ref003","doi-asserted-by":"crossref","first-page":"929","DOI":"10.1890\/0012-9658(1997)078[0929:CDICET]2.0.CO;2","article-title":"Compositional data in community ecology: the paradigm or peril of proportions?","volume":"78","author":"DA Jackson","year":"1997","journal-title":"Ecology"},{"key":"pcbi.1011240.ref004","doi-asserted-by":"crossref","first-page":"73","DOI":"10.1146\/annurev-statistics-010814-020351","article-title":"Microbiome, metagenomics, and high-dimensional compositional data analysis","volume":"2","author":"H Li","year":"2015","journal-title":"Annual Review of Statistics and Its Application"},{"issue":"2","key":"pcbi.1011240.ref005","doi-asserted-by":"crossref","first-page":"139","DOI":"10.1111\/j.2517-6161.1982.tb01195.x","article-title":"The statistical analysis of compositional data","volume":"44","author":"J Aitchison","year":"1982","journal-title":"Journal of the Royal Statistical Society: Series B (Methodological)"},{"issue":"3","key":"pcbi.1011240.ref006","doi-asserted-by":"crossref","first-page":"279","DOI":"10.1023\/A:1023818214614","article-title":"Isometric logratio transformations for compositional data analysis","volume":"35","author":"JJ Egozcue","year":"2003","journal-title":"Mathematical Geology"},{"issue":"1","key":"pcbi.1011240.ref007","doi-asserted-by":"crossref","first-page":"136","DOI":"10.1111\/j.2517-6161.1985.tb01341.x","article-title":"A general class of distributions on the simplex","volume":"47","author":"J Aitchison","year":"1985","journal-title":"Journal of the Royal Statistical Society: Series B (Methodological)"},{"key":"pcbi.1011240.ref008","unstructured":"Tsagris MT, Preston S, Wood AT. A data-based power transformation for compositional data; arXiv:1106.1451 [Preprint]. 2011. Available from: https:\/\/arxiv.org\/abs\/1106.1451."},{"issue":"1","key":"pcbi.1011240.ref009","doi-asserted-by":"crossref","first-page":"57","DOI":"10.1093\/biomet\/70.1.57","article-title":"Principal component analysis of compositional data","volume":"70","author":"J Aitchison","year":"1983","journal-title":"Biometrika"},{"issue":"4","key":"pcbi.1011240.ref010","first-page":"375","article-title":"Biplots of compositional data","volume":"51","author":"J Aitchison","year":"2002","journal-title":"Journal of the Royal Statistical Society: Series C (Applied Statistics)"},{"key":"pcbi.1011240.ref011","article-title":"Inferring correlation networks from genomic survey data","author":"J Friedman","year":"2012","journal-title":"Computational Biology"},{"issue":"2","key":"pcbi.1011240.ref012","doi-asserted-by":"crossref","first-page":"323","DOI":"10.1093\/biomet\/71.2.323","article-title":"Log contrast models for experiments with mixtures","volume":"71","author":"J Aitchison","year":"1984","journal-title":"Biometrika"},{"issue":"4","key":"pcbi.1011240.ref013","doi-asserted-by":"crossref","first-page":"785","DOI":"10.1093\/biomet\/asu031","article-title":"Variable selection in regression with compositional covariates","volume":"101","author":"W Lin","year":"2014","journal-title":"Biometrika"},{"issue":"2","key":"pcbi.1011240.ref014","first-page":"1019","article-title":"Regression analysis for microbiome compositional data","volume":"10","author":"P Shi","year":"2016","journal-title":"The Annals of Applied Statistics"},{"key":"pcbi.1011240.ref015","first-page":"1","article-title":"Regression models for compositional data: General log-contrast formulations, proximal optimization, and microbiome data applications","author":"PL Combettes","year":"2020","journal-title":"Statistics in Biosciences"},{"key":"pcbi.1011240.ref016","doi-asserted-by":"crossref","first-page":"2844","DOI":"10.21105\/joss.02844","article-title":"c-lasso\u2014a Python package for constrained sparse and robust regression and classification","volume":"6","author":"L Simpson","year":"2021","journal-title":"Journal of Open Source Software"},{"key":"pcbi.1011240.ref017","unstructured":"Ailer E, M\u00fcller CL, Kilbertus N. A causal view on compositional data; arXiv:2106.11234 [Preprint]. 2021. Available from: https:\/\/arxiv.org\/abs\/2106.11234."},{"issue":"5","key":"pcbi.1011240.ref018","doi-asserted-by":"crossref","first-page":"330","DOI":"10.1016\/j.annepidem.2016.03.002","article-title":"Compositional data analysis of the microbiome: fundamentals, tools, and challenges","volume":"26","author":"MC Tsilimigras","year":"2016","journal-title":"Annals of Epidemiology"},{"key":"pcbi.1011240.ref019","doi-asserted-by":"crossref","first-page":"2224","DOI":"10.3389\/fmicb.2017.02224","article-title":"Microbiome datasets are compositional: and this is not optional","volume":"8","author":"GB Gloor","year":"2017","journal-title":"Frontiers in Microbiology"},{"key":"pcbi.1011240.ref020","doi-asserted-by":"crossref","first-page":"2114","DOI":"10.3389\/fmicb.2017.02114","article-title":"Analysis of microbiome data in the presence of excess zeros","volume":"8","author":"A Kaul","year":"2017","journal-title":"Frontiers in Microbiology"},{"issue":"1","key":"pcbi.1011240.ref021","first-page":"1","article-title":"Analysis of microbial compositions: a review of normalization and differential abundance analysis","volume":"6","author":"H Lin","year":"2020","journal-title":"NPJ biofilms and microbiomes"},{"issue":"3","key":"pcbi.1011240.ref022","doi-asserted-by":"crossref","first-page":"253","DOI":"10.1023\/A:1023866030544","article-title":"Dealing with zeros and missing values in compositional data sets using nonparametric imputation","volume":"35","author":"JA Mart\u00edn-Fern\u00e1ndez","year":"2003","journal-title":"Mathematical Geology"},{"issue":"7","key":"pcbi.1011240.ref023","doi-asserted-by":"crossref","first-page":"e67019","DOI":"10.1371\/journal.pone.0067019","article-title":"ANOVA-like differential expression (ALDEx) analysis for mixed population RNA-Seq","volume":"8","author":"AD Fernandes","year":"2013","journal-title":"PloS one"},{"key":"pcbi.1011240.ref024","unstructured":"De La Cruz R, Kreft JU. Geometric mean extension for data sets with zeros; arXiv:1806.06403 [Preprint]. 2018. Available from: https:\/\/arxiv.org\/abs\/1806.06403."},{"key":"pcbi.1011240.ref025","unstructured":"Park J, Yoon C, Park C, Ahn J. Kernel Methods for Radial Transformed Compositional Data with Many Zeros. In: International Conference on Machine Learning. PMLR; 2022. p. 17458\u201317472."},{"issue":"7","key":"pcbi.1011240.ref026","doi-asserted-by":"crossref","first-page":"e1004977","DOI":"10.1371\/journal.pcbi.1004977","article-title":"Machine learning meta-analysis of large metagenomic datasets: tools and biological insights","volume":"12","author":"E Pasolli","year":"2016","journal-title":"PLoS computational biology"},{"issue":"7","key":"pcbi.1011240.ref027","doi-asserted-by":"crossref","first-page":"410","DOI":"10.1038\/s41579-018-0029-9","article-title":"Best practices for analysing microbiomes","volume":"16","author":"R Knight","year":"2018","journal-title":"Nature Reviews Microbiology"},{"key":"pcbi.1011240.ref028","doi-asserted-by":"crossref","first-page":"579","DOI":"10.3389\/fgene.2019.00579","article-title":"A review and tutorial of machine learning methods for microbiome host trait prediction","author":"YH Zhou","year":"2019","journal-title":"Frontiers in Genetics"},{"issue":"10","key":"pcbi.1011240.ref029","doi-asserted-by":"crossref","first-page":"635","DOI":"10.1038\/s41575-020-0327-3","article-title":"Gut microbiome, big data and machine learning to promote precision medicine for cancer","volume":"17","author":"G Cammarota","year":"2020","journal-title":"Nature Reviews Gastroenterology & Hepatology"},{"key":"pcbi.1011240.ref030","doi-asserted-by":"crossref","first-page":"191","DOI":"10.1007\/978-1-4614-7846-1_16","volume-title":"Topics in Applied Statistics","author":"J Chen","year":"2013"},{"issue":"1","key":"pcbi.1011240.ref031","doi-asserted-by":"crossref","first-page":"540","DOI":"10.1214\/17-AOAS1102","article-title":"Kernel-penalized regression for analysis of microbiome data","volume":"12","author":"TW Randolph","year":"2018","journal-title":"The Annals of Applied Statistics"},{"key":"pcbi.1011240.ref032","doi-asserted-by":"crossref","first-page":"60","DOI":"10.3389\/fmicb.2021.609048","article-title":"kernInt: A Kernel Framework for Integrating Supervised and Unsupervised Analyses in Spatio-Temporal Metagenomic Datasets","volume":"12","author":"E Ramon","year":"2021","journal-title":"Frontiers in Microbiology"},{"issue":"2","key":"pcbi.1011240.ref033","doi-asserted-by":"crossref","first-page":"113","DOI":"10.1177\/1471082X14535522","article-title":"Non-parametric regression for compositional data","volume":"15","author":"M Di Marzio","year":"2015","journal-title":"Statistical Modelling"},{"key":"pcbi.1011240.ref034","unstructured":"Tsagris M, Athineou G. Compositional: Compositional Data Analysis; 2021. Available from: https:\/\/CRAN.R-project.org\/package=Compositional."},{"issue":"5","key":"pcbi.1011240.ref035","doi-asserted-by":"crossref","first-page":"797","DOI":"10.1016\/j.ajhg.2015.04.003","article-title":"Testing in microbiome-profiling studies with MiRKAT, the microbiome regression-based kernel association test","volume":"96","author":"N Zhao","year":"2015","journal-title":"The American Journal of Human Genetics"},{"issue":"11","key":"pcbi.1011240.ref036","doi-asserted-by":"crossref","first-page":"1595","DOI":"10.1093\/bioinformatics\/btaa951","article-title":"MiRKAT: kernel machine regression-based global association tests for the microbiome","volume":"37","author":"N Wilson","year":"2021","journal-title":"Bioinformatics"},{"issue":"1","key":"pcbi.1011240.ref037","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1186\/s40168-022-01266-3","article-title":"Phylogeny-guided microbiome OTU-specific association test (POST)","volume":"10","author":"C Huang","year":"2022","journal-title":"Microbiome"},{"key":"pcbi.1011240.ref038","unstructured":"Li B, Ahn J. Reproducing Kernels and New Approaches in Compositional Data Analysis; arXiv:2205.01158 [Preprint]. 2022. Available from: https:\/\/arxiv.org\/abs\/2205.01158."},{"key":"pcbi.1011240.ref039","doi-asserted-by":"crossref","DOI":"10.1007\/978-3-030-28954-6","volume-title":"Explainable AI: interpreting, explaining and visualizing deep learning","author":"W Samek","year":"2019"},{"key":"pcbi.1011240.ref040","unstructured":"Molnar C. Interpretable machine learning. Lulu.com; 2020."},{"issue":"3","key":"pcbi.1011240.ref041","doi-asserted-by":"crossref","first-page":"e00434","DOI":"10.1128\/mBio.00434-20","article-title":"A framework for effective application of machine learning to microbiome-based classification problems","volume":"11","author":"BD Top\u00e7uo\u011flu","year":"2020","journal-title":"MBio"},{"issue":"2","key":"pcbi.1011240.ref042","doi-asserted-by":"crossref","first-page":"358","DOI":"10.2337\/dc20-1536","article-title":"Interpretable machine learning framework reveals robust gut microbiome features associated with type 2 diabetes","volume":"44","author":"W Gou","year":"2021","journal-title":"Diabetes Care"},{"issue":"12","key":"pcbi.1011240.ref043","doi-asserted-by":"crossref","first-page":"e1010714","DOI":"10.1371\/journal.pcbi.1010714","article-title":"Interpreting tree ensemble machine learning models with endoR","volume":"18","author":"A Ruaud","year":"2022","journal-title":"PLOS Computational Biology"},{"issue":"1","key":"pcbi.1011240.ref044","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1038\/s41598-021-93645-3","article-title":"Tree-aggregated predictive modeling of microbiome data","volume":"11","author":"J Bien","year":"2021","journal-title":"Scientific Reports"},{"key":"pcbi.1011240.ref045","first-page":"1189","article-title":"Greedy function approximation: a gradient boosting machine","author":"JH Friedman","year":"2001","journal-title":"Annals of Statistics"},{"issue":"2","key":"pcbi.1011240.ref046","doi-asserted-by":"crossref","first-page":"1","DOI":"10.17713\/ajs.v50i2.1069","article-title":"Impact of covariates in compositional models and simplicial derivatives","volume":"50","author":"J Morais","year":"2021","journal-title":"Austrian Journal of Statistics"},{"key":"pcbi.1011240.ref047","volume-title":"Learning with kernels: support vector machines, regularization, optimization, and beyond","author":"B Sch\u00f6lkopf","year":"2002"},{"issue":"7","key":"pcbi.1011240.ref048","article-title":"Universality, Characteristic Kernels and RKHS Embedding of Measures","volume":"12","author":"BK Sriperumbudur","year":"2011","journal-title":"Journal of Machine Learning Research"},{"key":"pcbi.1011240.ref049","unstructured":"Tops\u00f8e F. Jenson-Shannon divergence and norm-based measures of discrimination and variation; 2003. Available from: https:\/\/web.math.ku.dk\/~topsoe\/sh.ps."},{"key":"pcbi.1011240.ref050","first-page":"136","volume-title":"International Workshop on Artificial Intelligence and Statistics","author":"M Hein","year":"2005"},{"issue":"1","key":"pcbi.1011240.ref051","first-page":"129","article-title":"Diffusion kernels on statistical manifolds","volume":"6","author":"J Lafferty","year":"2005","journal-title":"Journal of Machine Learning Research"},{"issue":"12","key":"pcbi.1011240.ref052","doi-asserted-by":"crossref","first-page":"8228","DOI":"10.1128\/AEM.71.12.8228-8235.2005","article-title":"UniFrac: a new phylogenetic method for comparing microbial communities","volume":"71","author":"C Lozupone","year":"2005","journal-title":"Applied and Environmental Microbiology"},{"key":"pcbi.1011240.ref053","doi-asserted-by":"crossref","first-page":"477","DOI":"10.1890\/10-2402.1","article-title":"Measuring diversity: The importance of species similarity","volume":"93","author":"T Leinster","year":"2012","journal-title":"Ecology"},{"key":"pcbi.1011240.ref054","article-title":"Python 3 Reference Manual","author":"G van Rossum","year":"2009","journal-title":"CreateSpace"},{"key":"pcbi.1011240.ref055","author":"J Bradbury","year":"2018","journal-title":"JAX: composable transformations of Python+NumPy programs"},{"key":"pcbi.1011240.ref056","first-page":"2825","article-title":"Scikit-learn: Machine Learning in Python","volume":"12","author":"F Pedregosa","year":"2011","journal-title":"Journal of Machine Learning Research"},{"key":"pcbi.1011240.ref057","doi-asserted-by":"crossref","DOI":"10.1007\/978-1-4612-4380-9_16","volume-title":"Individual comparisons by ranking methods","author":"F Wilcoxon","year":"1992"},{"issue":"7516","key":"pcbi.1011240.ref058","doi-asserted-by":"crossref","first-page":"59","DOI":"10.1038\/nature13568","article-title":"Alterations of the human gut microbiome in liver cirrhosis","volume":"513","author":"N Qin","year":"2014","journal-title":"Nature"},{"issue":"1795","key":"pcbi.1011240.ref059","doi-asserted-by":"crossref","first-page":"20141988","DOI":"10.1098\/rspb.2014.1988","article-title":"Biogeographic patterns in below-ground diversity in New York City\u2019s Central Park are similar to those observed globally","volume":"281","author":"KS Ramirez","year":"2014","journal-title":"Proceedings of the Royal Society B: Biological Sciences"}],"updated-by":[{"DOI":"10.1371\/journal.pcbi.1011240","type":"new_version","label":"New version","source":"publisher","updated":{"date-parts":[[2023,7,13]],"date-time":"2023-07-13T00:00:00Z","timestamp":1689206400000}}],"container-title":["PLOS Computational Biology"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/dx.plos.org\/10.1371\/journal.pcbi.1011240","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2024,10,23]],"date-time":"2024-10-23T12:21:36Z","timestamp":1729686096000},"score":1,"resource":{"primary":{"URL":"https:\/\/dx.plos.org\/10.1371\/journal.pcbi.1011240"}},"subtitle":[],"editor":[{"given":"Luis Pedro","family":"Coelho","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"editor"}]}],"short-title":[],"issued":{"date-parts":[[2023,6,30]]},"references-count":59,"journal-issue":{"issue":"6","published-online":{"date-parts":[[2023,6,30]]}},"URL":"https:\/\/doi.org\/10.1371\/journal.pcbi.1011240","relation":{"new_version":[{"id-type":"doi","id":"10.1371\/journal.pcbi.1011240","asserted-by":"object"}]},"ISSN":["1553-7358"],"issn-type":[{"value":"1553-7358","type":"electronic"}],"subject":[],"published":{"date-parts":[[2023,6,30]]}}}