Genome-wide association studies (GWAS) have been remarkably successful in identifying susceptibility loci for complex diseases. These studies typically conduct single-variant tests of association to interrogate the genome in an agnostic fashion and, due to modest effect sizes, have come to rely on ever-greater sample sizes1,2 to make meaningful inferences. We have been less successful in developing methods that improve on existing simple approaches. In general, the genetic associations identified as genome-wide significant thus far account for only a modest proportion of variance in disease risk3. Indeed, there is now widespread recognition, if not consensus, that GWAS of disease susceptibility (for which, the relevant genetic effects may be very small) and pharmacologic traits (for which large effect sizes are not unusual)4,5 have resulted in limited conclusive findings on the genetic factors contributing to complex traits. Importantly, the functional significance of most discovered loci, including even those that have been the most reproducibly associated, remains unclear. Assigning a causal link to the nearest gene falls short of elucidating a functional connection, as recently demonstrated by the obesity-associated variants within FTO that form
