Prior to FST and PCA analyses, we LD-pruned the SNPs such that pairs with r2 > 0.1 within a window of 5 kb were pruned out using a step size of 50 bp. In addition, we excluded SNPs with MAF < 0.05. Pairwise FST estimates among the three ethnic groups within our study populations and among our study populations and the 1KGP3 reference populations were computed using smartpca. The tool computes the Hudson estimator of Wright’s FST which is immune to sample size differences, and uses a jackknife approach for ‘bootstrapping’, estimating a standard error and a Z score for the FST measurements. Regions of the genome with long-range LD listed in Table Table11 of Price et al.,86 were excluded from FST analysis and we further utilized smartpca’s LD regression function to correct for remnants of LD prior to FST estimation.
Do you have any questions about this protocol?
Post your question to gather feedback from the community. We will also invite the authors of this article to respond.