To prevent population stratification, the ancestry of each sample in the dataset is measured using STRUCTURE [31] or EIGENSTRAT [32] methods that compare genome-wide allele frequencies to those of HapMap ethnic groups. The results of these analyses can be used to either exclude samples with similarity to a non-target population, or they can be used as a covariate in association analysis. EIGENSTRAT is commonly used in this circumstance, where principle component analysis is used to generate principle component values that could be described as an “ethnicity score”. When used as covariates, these scores adjust for minute ancestry effects in the data.
