The power of numbers

The technical and methodological advancements, as well as the knowledge accrued over the past decade on the haplotype block structure of the human genome, have enabled investigators to tackle the complexity of the genetic architecture of type 2 diabetes in populations of European and non-European descent by performing large-scale genome-wide association studies (GWAS) for both common and rare genetic variants. Interestingly, while interpreting the GWAS results one may observe that as the number of identified type 2 diabetes risk variants has increased over time, and the loci uncovered by earlier GWAS have been further replicated in larger association studies, the individual (per-allele) effect estimate has become smaller than the one originally detected in the discovery GWAS. This may be due to the non-mutually exclusive occurrence of two statistical phenomena, usually dubbed as "winner's curse" and "spectrum bias" effects. The present commentary discusses the work of the China Kadoorie Biobank Collaborative Group, which sought to provide a demonstration of the calculation of (relatively) unbiased allelic effect sizes for a set of 56 established type 2 diabetes risk variants in a large population-based cohort study of Chinese adult individuals. In particular we critically discuss whether theGWAS approach should remain a matter of statistical constraints only, or whether its integration with functional maps may highlight some sub-threshold loci as informative as those that reach genome-wide significance. The complementary information that could arise from the full integration of the genetic and functional maps holds the promise of potentially uncovering clinically relevant mechanistic insights and might expand the regulatory framework in which to interpret the functional follow-up and fine-mapping currently ongoing at established type 2 diabetes risk loci.