Strategies for genome-wide association studies: optimization of study designs by the stepwise focusing method

Abstract

Recently, the use of genome-wide linkage disequilibrium (LD) analysis to localize traits has attracted much attention because of the introduction of high-throughput genotyping systems. However, a limitation of such studies is often the total cost of genotyping in addition to sample size. Therefore, it is important to estimate optimal conditions for such a study given the total cost of genotyping. In the present study, we have introduced the "stepwise focusing method," in which candidate markers are selected in a stepwise fashion. In the first focusing step, samples from both case and control groups are genotyped at a certain number of single-nucleotide polymorphisms (SNPs) (for example, 50000), and the markers that exhibit significant intergroup differences by a chi(2) test are selected. In the first step, the risk of type I error is set rather high (for example, 0.1), and, therefore, most of the selected markers are false positives. In the second step, the markers selected in the first step are tested by using samples obtained from a different set of case-control samples. We performed extensive simulation studies to estimate both the type I error and the power of the test by changing parameters such as genotype relative risk, disease allele frequency, and sample size. If the total number of genotypings was limited, the stepwise focusing method yielded optimal conditions and was more powerful than conventional methods.