With more multi-ancestral biobanks and large consortia becoming available, including future releases of data from MVP, the Global Biobank Meta-analysis Initiative57 and the All of Us Research Program58, we anticipate that the gap between findings in EUR and other populations will diminish. Confounding effects, including socioeconomic status, may bias our results; the rg with EA is −0.21 (P = 7.57 × 10−31), indicating a shared genetic architecture between PAU and EA, a socioeconomic factor that influences many psychiatric traits (and nonpsychiatric traits as well)31. Genetic nurture, or indirect genetic effects—effects of alleles in parents on offspring through the environment—exist in many GWAS59. Imputation of parental genotypes using family data could improve estimates of direct genetic effects for PAU60. We note that the current findings are not sufficient for clinical risk prediction at the individual level, given the limited SNP-based heritability and small proportion of variance explained by PRS.
