First, ADMIXTURE was run on 30 random 100,000-SNP subsets of the core SNP set with k (the
number of groups) ranging from 5 to 18, and k = 9 was chosen because it was the minimal value of k to separate all
previously known groups (cA, cB, XI, GJ-trp, GJ-tmp and part of GJ-sbtrp).
With k = 9, ADMIXTURE was then run again on the whole core SNP set nine times with varying random seeds;
the Q-matrices were aligned using CLUMPP
software and clustered on the basis of similarity.
Then, the matrices belonging to the largest cluster were averaged to produce the final matrix of
Finally, the group membership for each sample was defined by applying the threshold of ≥ 0.65 to this
matrix. Samples with admixture components <0.65 were classified as follows. If the sum of components for subpopulations within
the major groups XI and GJ was ≥ 0.65, the samples were classified as XI-adm or GJ-adm, respectively, and the remaining samples
were deemed 'fully' admixed (admix).
The phylogenetic tree is displayed below, branches of which are coloured according to the K9 classification.