{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,3,26]],"date-time":"2026-03-26T12:11:20Z","timestamp":1774527080185,"version":"3.50.1"},"reference-count":25,"publisher":"Oxford University Press (OUP)","issue":"1","license":[{"start":{"date-parts":[[2025,11,9]],"date-time":"2025-11-09T00:00:00Z","timestamp":1762646400000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/academic.oup.com\/pages\/standard-publication-reuse-rights"}],"funder":[{"name":"Office of Naval Research, National Science Foundation"},{"name":"National Institutes of Health of the United States"},{"DOI":"10.13039\/501100000038","name":"Natural Sciences and Engineering Research Council of Canada","doi-asserted-by":"publisher","id":[{"id":"10.13039\/501100000038","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2026,2,13]]},"abstract":"<jats:title>Summary<\/jats:title>\n                  <jats:p>Recent work has explored data thinning, a generalization of sample splitting that involves decomposing a (possibly matrix-valued) random variable into independent components. In the special case of an $ n\\times p $ random matrix with independent and identically distributed $ N_{p}(\\mu,\\Sigma) $ rows, Dharamshi et\u00a0al. (2026)provided a comprehensive analysis of the settings in which thinning is or is not possible: briefly, if $ \\Sigma $ is unknown then one can thin provided that $ n \\gt 1 $. However, in some situations a data analyst may have access only to summary statistics of the data, e.g., due to privacy considerations. While the sample mean follows a Gaussian distribution, the sample covariance follows, up to scaling, a Wishart distribution, for which no thinning strategies have yet been proposed. In this note, we fill this gap: we show that it is possible to generate two or more independent data matrices with independent $ N_{p}(\\mu,\\Sigma) $ rows, based only on the sample mean and sample covariance matrix. These independent data matrices can either be used directly within a train-test paradigm or be used to derive independent summary statistics. Furthermore, they can be recombined to yield the original sample mean and sample covariance. The key insight that enables this development is an algorithm that decomposes a Wishart random matrix into a matrix square root with independent and identically distributed Gaussian rows.<\/jats:p>","DOI":"10.1093\/biomet\/asaf081","type":"journal-article","created":{"date-parts":[[2025,11,9]],"date-time":"2025-11-09T20:35:31Z","timestamp":1762720531000},"source":"Crossref","is-referenced-by-count":0,"title":["Thinning a Wishart random matrix"],"prefix":"10.1093","volume":"113","author":[{"given":"A","family":"Dharamshi","sequence":"first","affiliation":[{"name":"Department of Biostatistics, University of Washington , Seattle, Washington 98195,","place":["U.S.A"]}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"A","family":"Neufeld","sequence":"additional","affiliation":[{"name":"Department of Mathematics and Statistics, Williams College , 18 Hoxsey Street , Williamstown, Massachusetts 01267, . acn2@williams.edu","place":["U.S.A"]}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"L L","family":"Gao","sequence":"additional","affiliation":[{"name":"Department of Statistics, University of British Columbia , Vancouver, British Columbia V6T 1Z4, \u00a0 lucy.gao@ubc.ca","place":["Canada"]}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"D","family":"Witten","sequence":"additional","affiliation":[{"name":"Departments of Biostatistics and Statistics, University of Washington , Seattle, Washington 98195, . dwitten@uw.edu","place":["U.S.A"]}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"J","family":"Bien","sequence":"additional","affiliation":[{"name":"Department of Data Sciences and Operations, University of Southern California , 3670 Trousdale Parkway, Los Angeles, California 90007, . jbien@usc.edu","place":["U.S.A"]}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"286","published-online":{"date-parts":[[2025,11,9]]},"reference":[{"key":"2026032607090939300_asaf081-B1","volume-title":"An Introduction to Multivariate Statistical Analysis","author":"Anderson","year":"2003","edition":"3rd ed"},{"key":"2026032607090939300_asaf081-B2","first-page":"485","article-title":"Model selection through sparse maximum likelihood estimation for multivariate Gaussian or binary data","volume":"9","author":"Banerjee","year":"2008","journal-title":"J. Mach. Learn. Res"},{"key":"2026032607090939300_asaf081-B3","author":"Borchers","year":"2023"},{"key":"2026032607090939300_asaf081-B4","doi-asserted-by":"crossref","first-page":"304","DOI":"10.1038\/nn.4499","article-title":"Computational approaches to fMRI analysis","volume":"20","author":"Cohen","year":"2017","journal-title":"Nature Neurosci."},{"key":"2026032607090939300_asaf081-B5","doi-asserted-by":"crossref","first-page":"441","DOI":"10.1093\/biomet\/62.2.441","article-title":"A note on data-splitting for the evaluation of significance levels","volume":"62","author":"Cox","year":"1975","journal-title":"Biometrika"},{"key":"2026032607090939300_asaf081-B6","doi-asserted-by":"crossref","first-page":"asaf057","DOI":"10.1093\/biomet\/asaf057","article-title":"Decomposing Gaussians with unknown covariance","volume":"113","author":"Dharamshi","year":"2026","journal-title":"Biometrika"},{"key":"2026032607090939300_asaf081-B7","doi-asserted-by":"crossref","first-page":"511","DOI":"10.1080\/01621459.2024.2353948","article-title":"Generalized data thinning using sufficient statistics","volume":"120","author":"Dharamshi","year":"2025","journal-title":"J. Am. Statist. Assoc"},{"key":"2026032607090939300_asaf081-B8","doi-asserted-by":"crossref","first-page":"432","DOI":"10.1093\/biostatistics\/kxm045","article-title":"Sparse inverse covariance estimation with the graphical lasso","volume":"9","author":"Friedman","year":"2008","journal-title":"Biostatistics"},{"key":"2026032607090939300_asaf081-B9","doi-asserted-by":"crossref","DOI":"10.1201\/9780203749289","volume-title":"Matrix Variate Distributions","author":"Gupta","year":"2018"},{"key":"2026032607090939300_asaf081-B10","first-page":"311","article-title":"A note on Mitra\u2019s paper A density free approach to the matrix variate beta distribution","volume":"32","author":"Khatri","year":"1970","journal-title":"Sankhy\u0101 A"},{"key":"2026032607090939300_asaf081-B11","doi-asserted-by":"crossref","first-page":"135","DOI":"10.1080\/01621459.2023.2270748","article-title":"Data fission: splitting a single data point","volume":"120","author":"Leiner","year":"2025","journal-title":"J. Am. Statist. Assoc"},{"key":"2026032607090939300_asaf081-B12","doi-asserted-by":"crossref","first-page":"451","DOI":"10.1093\/biomet\/92.2.451","article-title":"Monte Carlo conditioning on a sufficient statistic","volume":"92","author":"Lindqvist","year":"2005","journal-title":"Biometrika"},{"key":"2026032607090939300_asaf081-B13","author":"Mezzadri","year":"2007"},{"key":"2026032607090939300_asaf081-B14","first-page":"81","article-title":"A density-free approach to the matrix variate beta distribution","volume":"32","author":"Mitra","year":"1970","journal-title":"Sankhy\u0101"},{"key":"2026032607090939300_asaf081-B15","volume-title":"Aspects of Multivariate Statistical Theory","author":"Muirhead","year":"2009"},{"key":"2026032607090939300_asaf081-B16","first-page":"1","article-title":"Data thinning for convolution-closed distributions","volume":"25","author":"Neufeld","year":"2024","journal-title":"J. Mach. Learn. Res"},{"key":"2026032607090939300_asaf081-B17","doi-asserted-by":"crossref","first-page":"151","DOI":"10.1080\/01621459.2024.2421998","article-title":"Discussion of \u2018Data fission: splitting a single data point\u2019","volume":"120","author":"Neufeld","year":"2025","journal-title":"J. Am. Statist. Assoc"},{"key":"2026032607090939300_asaf081-B18","doi-asserted-by":"crossref","first-page":"117","DOI":"10.1038\/nrg.2016.142","article-title":"Dissecting the genetics of complex traits using summary association statistics","volume":"18","author":"Pasaniuc","year":"2017","journal-title":"Nature Rev. Genet"},{"key":"2026032607090939300_asaf081-B19","doi-asserted-by":"crossref","first-page":"597","DOI":"10.1093\/biomet\/asac070","article-title":"Splitting strategies for post-selection inference","volume":"110","author":"Rasines","year":"2023","journal-title":"Biometrika"},{"key":"2026032607090939300_asaf081-B20","doi-asserted-by":"crossref","first-page":"229","DOI":"10.1214\/009053605000000877","article-title":"Adaptive nonparametric confidence sets","volume":"34","author":"Robins","year":"2006","journal-title":"Ann. Statist."},{"key":"2026032607090939300_asaf081-B21","doi-asserted-by":"crossref","first-page":"494","DOI":"10.1214\/08-EJS176","article-title":"Sparse permutation invariant covariance estimation","volume":"2","author":"Rothman","year":"2008","journal-title":"Electron. J. Statist."},{"key":"2026032607090939300_asaf081-B22","doi-asserted-by":"crossref","first-page":"1537","DOI":"10.1214\/aos\/1065705118","article-title":"Singular Wishart and multivariate beta distributions","volume":"31","author":"Srivastava","year":"2003","journal-title":"Ann. Statist."},{"key":"2026032607090939300_asaf081-B23","doi-asserted-by":"crossref","first-page":"679","DOI":"10.1214\/17-AOS1564","article-title":"Selective inference with a randomized response","volume":"46","author":"Tian","year":"2018","journal-title":"Ann. Statist."},{"key":"2026032607090939300_asaf081-B24","doi-asserted-by":"crossref","first-page":"401","DOI":"10.1007\/BF02288916","article-title":"Multidimensional scaling: I. Theory and method","volume":"17","author":"Torgerson","year":"1952","journal-title":"Psychometrika"},{"key":"2026032607090939300_asaf081-B25","doi-asserted-by":"crossref","first-page":"19","DOI":"10.1093\/biomet\/asm018","article-title":"Model selection and estimation in the Gaussian graphical model","volume":"94","author":"Yuan","year":"2007","journal-title":"Biometrika"}],"container-title":["Biometrika"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/academic.oup.com\/biomet\/advance-article-pdf\/doi\/10.1093\/biomet\/asaf081\/65248516\/asaf081.pdf","content-type":"application\/pdf","content-version":"am","intended-application":"syndication"},{"URL":"https:\/\/academic.oup.com\/biomet\/article-pdf\/113\/1\/asaf081\/65248516\/asaf081.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/academic.oup.com\/biomet\/article-pdf\/113\/1\/asaf081\/65248516\/asaf081.pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,3,26]],"date-time":"2026-03-26T11:09:14Z","timestamp":1774523354000},"score":1,"resource":{"primary":{"URL":"https:\/\/academic.oup.com\/biomet\/article\/doi\/10.1093\/biomet\/asaf081\/8317439"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,11,9]]},"references-count":25,"journal-issue":{"issue":"1","published-print":{"date-parts":[[2026,2,13]]}},"URL":"https:\/\/doi.org\/10.1093\/biomet\/asaf081","relation":{},"ISSN":["0006-3444","1464-3510"],"issn-type":[{"value":"0006-3444","type":"print"},{"value":"1464-3510","type":"electronic"}],"subject":[],"published-other":{"date-parts":[[2026]]},"published":{"date-parts":[[2025,11,9]]},"article-number":"asaf081"}}