Replication and heritability of prostate cancer risk variants: impact of population-specific factors.

BACKGROUND Prostate cancer incidence and mortality rates vary across populations, with African American men exhibiting the highest rates. To date, genome-wide association studies have identified 104 SNPs independently associated with prostate cancer in men of European ancestry. METHODS We investigated whether the ability to replicate findings for these 104 SNPs in African American, Asian, and Latino populations depends on variation in risk allele frequencies (RAF), strength of associations, and/or patterns of linkage disequilibrium (LD) at the associated loci. We extracted estimates of effect from the literature, and determined RAF and LD information across the populations from the 1000 Genomes Project. RESULTS Risk variants were largely replicated across populations. Relative to Europeans, 83% had smaller effect sizes among African Americans and 73% demonstrated smaller effect sizes among Latinos. Among Asians, however, 56% showed larger effect sizes than among Europeans. The largest difference in RAFs was observed between European and African ancestry populations, but this difference did not impact our ability to replicate. The extent of LD within 250 kb of risk loci in Asian ancestry populations was suggestively lower for variants that did not replicate (P = 0.013). CONCLUSIONS Despite substantial overlap in prostate cancer risk SNPs across populations, the variation in prostate cancer incidence among different populations may still in part reflect unique underlying genetic architectures. IMPACT Studying different ancestral populations is crucial for deciphering the genetic basis of prostate cancer.