We next identified studies that contained valid comparisons of the performance of polygenic scores in European ancestry participants and at least one other ancestry. Specifically, matched analyses (from two or more ancestry groups, from any given publication) had to use the same genotyping chip for all samples, the same weights for variants, the same algorithm for constructing polygenic scores, and the same methods of measuring phenotypes across all participants. Results from 26 studies met inclusion criteria, and there was minimal overlap in samples used (see Supplementary Data 2). From these studies we then extracted effect size metrics for each ancestry group. Effect size, refers to variance explained (the preferred metric) and to beta coefficients from regression, odds ratio (OR), or area under the curve (AUC), as available. Score performance for each analysis of a sample of non-European ancestry, was normalized to performance in the matched European ancestry sample from the same study, by dividing effect sizes (i.e., non-European ancestry/European ancestry). Odds ratios were first converted to log(OR). We multiplied values by 100 so that performance for each non-European ancestry sample
