Statistical methodology has been developed and implemented into software to aid in detecting and adjusting for population stratification in GWAS. Genomic control [28,29], aims to control for population stratification by first estimating an inflation factor, then adjusting all of the test statistics downward by this factor. Several variations on genomic control have been developed, and a recent review and critical evaluation of genomic control methods [30] recommended genomic control F (GCF) [31] as the most appropriate variation. GCF does not assume the inflation factor is measured without error, and refines this factor accordingly. Structured association [32], implemented in the STRUCTURE software [33] uses genotype data to infer population structure and subsequently performs tests of association within each inferred subpopulation. STRUCTURE may also be used to identify individual samples that do not cluster with the majority of the samples. These samples can then be eliminated from the analysis.
