An important finding of our work that the inclusion of variants from an African ancestry population results in less disparity in PRS accuracy across other populations illustrates the need to recruit diverse populations in GWASs and make these data readily available. Large consortia such as H3Africa, PAGE, the Million Veterans Program, and All of Us are undertaking efforts to support this initiative. Based on our analysis of HbA1c, asthma, and prostate cancer in the UK Biobank, we find that improvement in PRS prediction accuracy is currently possible by incorporating findings from GWASs in African ancestry populations, albeit with lower power. In addition to smaller sample sizes, this potential improvement may be limited by ascertainment bias in what SNPs are included on genotyping arrays and poorer imputation in non-Europeans. GWAS arrays and their imputation have substantially higher coverage among Europeans, and this may result in decreased PRS portability of findings across traits; in such situations, whole-genome sequencing in diverse populations may be needed in order to improve accuracy.27,28 Our study and others support the immense genetic diversity that can be unlocked
