How to interpret a genome-wide association study (GWAS)?

Genome-wide association studies (GWAS) aim at identifying genetic susceptibility to multifactorial diseases. They compare the frequency of several hundred thousand genetic variants distributed across the chromosomes in a group of cases with a given disease and a group of controls, using high-throughput genotyping technologies. In contrast with candidate gene association studies, GWAS use an agnostic approach, requiring no a priori hypothesis about the genes involved. The important number of statistical tests performed most often requires access to computer clusters for adequate processing power, and correction for multiple testing needs to be performed, a p-value < 5 x 10(-8) being usually considered as statistically significant. Large samples are needed to reach sufficient statistical power, thus requiring multicenter projects led by international consortia. It is important to take into account the ethnic and geographic origin of study participants, in order to avoid false positive associations due to population stratification. Another crucial point, as for any genetic association study, is to replicate significant associations in an independent population. Over the past years, GWAS have lead to the identification of hundreds of novel genetic variants associated with various multifactorial diseases. Interestingly these were generally located within or close to previously unsuspected genes. Discovering new susceptibility genes is essential to improve our understanding of the biological pathways involved in multifactorial diseases. This could help identify new therapeutic targets and strategies. Another potential application is improved risk prediction and personalized medicine or therapy. So far, GWAS have been mainly focused on common single nucleotide polymorphisms, i.e. with a relatively high minor allele frequency. Other types of genetic variation are likely to contribute substantially to the heritability of multifactorial diseases.