Polygenic scores can be constructed for any complex genetic phenotype for which appropriate GWAS (or other robust association) results are available. The challenges inherent in using polygenic scores—including modest capacity for prediction and necessary considerations regarding statistical power—have been reviewed previously21,26. Recent research has focused on the generalizability of polygenic scores to non-European ancestry populations27. There is good reason to anticipate reduced predictive power in non-European ancestry samples because of differences in variant frequencies and linkage disequilibrium patterns between populations12,28. However, the magnitude of performance decrease is largely unknown. Few systematic studies of polygenic score performance across different ancestry groups are available, though see Hoffman et al.13 for an investigation of blood pressure metrics and a recent UK Biobank analysis27. Further, some previous findings may need to be re-evaluated in light of newer findings about relationships between ancestry and GWAS results29–31. Thus, there is a need for systematic evaluation of polygenic score performance across multiple populations, phenotypes, and samples.
