 Abstract polygenic scores, PGSs, which are individual-level measures of genetic predispositions to certain traits, have been developed primarily using European ancestry samples. However, these scores may be less accurate in non-European ancestry individuals due to differences in allele frequencies between different ethnicities. In this study, researchers investigated the effects of sample size and ancestry composition on PGS performance for 15 traits in UK Biobank. They found that for some traits, PGS estimates, based on a smaller African ancestry training set performed better than those based on a larger European ancestry, only training set. This suggests that targeted data collection from underrepresented groups can help to reduce existing disparities in PGS performance. This article was authored by Priyuk Lehman, Maxine McIntosh, Gil McVean, and others. We are article.tv, links in the description below.