Analysis of GWAS: Opportunities and Obstacles

Genome-wide association studies (GWAS) have discovered hundreds of loci associated with scores of complex diseases and traits. However, the loci identified to date typically explain only a small fraction of variation in a trait--far smaller than the fraction that family and twin studies suggest is due to genetics--an observation which has been called the problem of missing heritability.

"Missing heritability" is to a large extent due to one of the main obstacles facing GWAS: very large sample sizes are needed to reliably identify modest but real associations between common markers and complex traits. I present theoretical and empirical examples illustrating this point. Moreover, to date GWAS have adopted a simple analytic approach, focused on the marginal effects of individual SNPs (i.e. analyzing each SNP separately, effectively estimating an average effect over genetic and environmental variation). Approaches that leverage gene-gene or gene-environment interaction may identify markers that simpler analytic approaches miss. However, incorporating environmental risk factors requires renewed attention to issues of study design and exposure measurement. In particular, the benefits of studies of gene-environment interaction will be maximized by broadening the range of studied exposures, for example by conducting GWAS in new populations or through international collaborations.

The hundreds of trait loci identified by GWAS to date represent only the initial returns on the investment in these studies. Future studies and additional analysis of existing studies will likely exponentially increase the utility of studies completed to date.