Our results have several important implications for human disease gene mapping and personal genomics. In particular, the vast majority of protein-coding variation is evolutionarily recent, rare, and enriched for deleterious alleles. Thus, rare variation likely makes an important contribution to human phenotypic variation and disease susceptibility. However, detecting the effects of rare variants requires very large sample sizes, because the power to detect an association is low for most human genes. Accounting for the SFS on a gene-by-gene basis should facilitate the development of more powerful association tests. Additionally, because most rare SNVs are population-specific, replication of disease associations across populations may be challenging. Lastly, as whole-genome sequencing at high coverage becomes increasingly feasible, statistical and experimental methods that accurately identify functionally important protein-coding and regulatory variation are needed to empower association studies, identify variants causally related to disease, and provide clinically actionable information.
