The Effect of Population Structure on Murine Genome-Wide Association Studies

Abstract

Population structure (PS) has been shown to cause false positive signals in genome-wide association studies (GWAS). Since PS correction is routinely used in human GWAS, it was assumed that it also should be utilized for murine GWAS using inbred strains. Nevertheless, there are fundamental differences between murine and human GWAS, and the impact of PS on murine GWAS results has not been thoroughly investigated. To assess the impact of PS on murine GWAS, we examined 8223 datasets that characterized biomedical responses in panels of inbred mouse strains. Surprisingly, we found that the PS had a minimal impact on datasets characterizing responses in [≤]20 strains; and relatively little impact on the majority of datasets characterizing >20 strains. Moreover, there were examples where association signals within known causative genes could be rejected if PS correction methods were utilized. PS assessment should be carefully used and should be considered in conjunction with other criteria when assessing the candidate genes identified in GWAS using inbred mouse strains.