for the ith and jth samples, where Nij is the total number of SNPs with genotype data available for both samples. Note that for GWA genotyping platforms, suitably thinned subsets of SNPs should be used to minimize the effects of linkage disequilibrium (LD) between markers (e.g. SNPs with r2 <0.2 between each other). IBS can take values between 0 and 1, where 1 represents complete identity between a pair of samples. Classical multidimensional scaling is then applied to the matrix of distances, 1-IBS, to identify axes of genetic variation between samples. The resulting scores for the ith individual on these axes are denoted as xi1,xi2,…,xiT, where xit corresponds to the axis with the tth largest eigenvalue.
