Guidelines for Evaluating the Comparability of Down-Sampled GWAS Summary Statistics

• DOI: 10.1007/s10519-023-10152-z
• Publication Date: 2023-09-15T12:01:54Z
• Authors: Camille M. Williams; Holly Poore; Peter T. Tanksley; Hyeokmoon Kweon; Natasia S. Courch...; Diego Londono-Correa; Travis T. Mallard; Peter Barr; Philipp D. Koelli...; Irwin D. Waldman; Sandra Sanchez-Roige; K. Paige Harden; Abraham A. Palmer; Danielle M. Dick; Richard Karlsson ...
• Keywords: Comparability; Genome-wide Association Study; Health psychology

Proprietary genetic datasets are valuable for boosting the statistical power of genome-wide association studies (GWASs), but their use can restrict investigators from publicly sharing resulting summary statistics. Although researchers resort to down-sampled versions that exclude restricted data, down-sampling reduces and might change etiology phenotype being studied. These problems further complicated when using multivariate GWAS methods, such as genomic structural equation modeling (Genomic SEM), model correlations across multiple traits. Here, we propose a systematic approach assess comparability statistics include versus data. Illustrating this with an externalizing factor, assessed impact on (1) strength signal in univariate GWASs, (2) factor loadings fit Genomic SEM, (3) at level, (4) insights gene-property analyses, (5) pattern other traits, (6) polygenic score analyses independent samples. For GWAS, although resulted loss fewer significant loci; fit, correlations, were found robust. Given importance data advancement open science, recommend who generate share report these accompanying documentation support researchers'