Commentary: Genetic association studies see light at the end of the tunnel

In this month’s issue Ioannidis et al. provides a welcome guide to interpreting data from genetic association studies. The authors’ efforts are important for two main reasons. First, genetic associations have been fraught with difficulty over the past 10 years in their attempts to uncover DNA polymorphisms that alter disease risk. The vast majority of reported associations, typically between a single nucleotide polymorphism (SNP) and a disease, were not replicated. The reasons for this are now well understood and have been discussed before. The main problem is that geneticists have several 100 000 risk factors to study in the form of common polymorphisms but only a few are likely to be involved in any one disease. This makes the prior odds that any one variant is associated very low and therefore very stringent P-values are needed to provide any confidence in the statistical evidence. Second, genome wide association (GWA) studies, that test several 100 000 DNA variants in a single experiment, have arrived in abundance in early 2007 making it potentially even harder for epidemiologists to pick their way through the data to decide what is real.