High hopes have been placed on the promise of genome-wide association studies in revealing the genetic components of complex diseases. New research has brought the potential of such studies to the fore, identifying four novel loci that are associated with type 2 diabetes mellitus.

Because of the well-established heritability of type 2 diabetes, the search for causative variants has been intense. Despite this, most of the variants implicated so far confer only a small proportion of the overall risk, the most convincing association being with a variant in the gene that encodes the transcription factor TCF7L2.

Using high-density genotyping arrays, the authors took a two-stage approach to identifying other diabetes-associated variants. The first round tested 392,935 SNPs in a French case–control cohort in which subjects were limited to those with affected relatives and normal body mass index, to exclude obesity-related risk alleles. This stage yielded 59 SNPs with a strong association for the disease. The authors then homed in on the SNPs with the greatest risk association through a second round of analysis, this time using a larger cohort in which the inclusion criteria were relaxed to maximize the power of detecting an association. Following both stages of analysis, eight SNPs corresponding to five unique loci were identified as risk variants, with one of these loci confirming the known TCF7L2 association.

Of the seven previously unidentified SNPs, the most significant association corresponded to a non-synonymous polymorphism in the SLC30A8 gene, which encodes a zinc transporter that is expressed only in β-cells, and is therefore likely to be involved in insulin production. Two other SNPs were mapped to a linkage-disequilibrium block containing genes that encode insulin-degrading enzyme (IDE), the homeodomain protein HHEX, which is essential for hepatic and pancreatic development, and a kinesin-interacting factor, KIF11, although further study is required to pinpoint the variant that confers the risk. Less significantly associated loci contain genes in the hedgehog and WNT pathways, which have roles in pancreatic development and the regulation of insulin secretion.

For three of the four novel loci, the risk allele is the major allele, so their value for familial clustering and individual risk assessment is limited. However, the authors observed little epistasis between individual loci, suggesting that they, along with the TCF7L2 variant, could account for as much as 70% of the population-attributable risk for type 2 diabetes. Interestingly, seven of the eight identified SNPs also represent the ancestral allele, supporting the idea that the adaptations of our ancestors might be detrimental in the modern environment, helping to explain the increasing prevalence of complex disease.

These results could represent the first steps towards a greater understanding of the pleiotropic mechanisms behind type 2 diabetes, and further work with larger, more diverse cohorts is expected to uncover more risk-carrying alleles. No doubt, genome-wide studies into other complex diseases will offer up their own insights in the near future.