Recovering unused information in genome-wide association studies: the benefit of analyzing SNPs out of Hardy-Weinberg equilibrium.

摘要: Although the rapid advancements in high throughput genotyping technology have made genome-wide association studies possible, these remain an expensive undertaking, especially when considering large sample sizes necessary to find small moderate effect that define complex diseases. It is therefore prudent utilize all possible information contained a scan. We propose straightforward analytical approach tests often unused SNP data without sacrificing statistical validity. simulate genotype miscalls under variety of models consistent with observed miscall rates and test for departures from HWE using standard Pearson's χ2-test. true disease susceptibility loci subjected various patterns can be largely out and, thus, candidates removal before testing. These loci, we demonstrate, maintain sufficient power even extreme error models. additionally show random null SNPs, independent phenotype, do not induce bias case–control or cohort studies, suggest significant should prevent being tested conducting scenarios. However, findings SNPs are must treated more carefully than ‘regular' findings, example, by re-genotyping same study different technology.