Alternative polygenic associations

Abstract

Human Genetics To predict traits and/or disease from genomic information, we must understand the genetics of the associations between genotype and phenotype. Polygenic risk scores are increasingly used to make predictions of phenotype on the basis of all the loci that correlate with a specific trait identified in genome-wide association studies (GWASs). However, population-level genetics and variable environmental effects can affect GWAS results. To determine the extent of any skewed GWAS results, Mostafavi et al. examined the accuracy of GWASs in a relatively homogeneous population: individuals of European-British descent within the UK Biobank. They found that sex, age, and socioeconomic status can affect GWAS signals. The authors suggest that using data from families within GWASs should help pull out direct and indirect effects, such as shared environmental factors and assortative mating to increase the accuracy of polygenic risk score prediction. eLife 9 , e48376 (2020).