TagSNP selection based on pairwise LD criteria and power analysis in association studies

TagSNP selection is an important step in designing case control association studies. Among selection methods that have proliferated, the ones based on pairwise LD measurement are attractive for the purpose of designing association studies. The goal is to minimize the number of markers selected for genotyping in a particular platform and therefore reduce genotyping cost while simultaneously representing information provided by all other markers. Depending on the platform, it is also important to select sets that are robust against occasional genotyping failure. An array of methods has been proposed to effectively select these tagSNPs using various criteria. In this study, we extend the algorithms used in FESTA, a computer program we previously developed for picking tagSNPs using r2 criteria. We applied FESTA to the HapMap whole chromosome data in two different populations, and we also performed a power analysis for case-control association studies using simulated data. FESTA chooses 294322 tagSNPs in the autosomes in the CEPH samples. The YORUBA samples require 61.5% more tagSNPs than the CEPH samples. The power study showed that limiting ourselves to only tagSNPs, instead of choosing all SNPs in the interval for an association study, results in a power loss of only about 5-10%.