Fig. 5: Bayesian population assignment for four clusters (K=4): Admixture plots are histograms of admixture coefficients and show the relative assignment of the individuals to the different relevant clusters. On the y-axis, the proportion of admixture per individual is shown. The x-axis spans the entire dataset with the current percentage of allocation to a cluster. Each bar represents one individual. A one-colored bar belongs exclusively to one cluster, whereas a multi-colored bar indicates an affiliation to different clusters. The individuals on the x-axis are sorted by longitude from west to east: blue=Western Siberia, purple=Western Yakutia, yellow=Eastern Yakutia, green=Chukotka.
3.3 Biogeographic inference - Approximate Bayesian Computation
Since linkage disequilibrium (LD) can affect biogeographic inference, SNPs with strong LD were pruned. In order to minimize sequencing and assembly errors the SNPs showing severe distortion of the HWE (p < 0.05), MAF lower than 5%, and MISS above 20% were filtered out. Thereby, the number of SNPs was reduced to 2733.
Table 1: Posterior probability and confidence interval (CI) for every DIYABC scenario
Scenario Posterior probability (95%CI)
1 0.0016 (0.0000-0.6450)
2 0.0001 (0.0000-0.6451)
3 0.0068 (0.0000-0.6448)
4 0.0000 (0.0000-0.6451)
5 0.0000 (0.0000-0.6451)
6 0.0001 (0.0000-0.6451)
7 0.2347 (0.0000-0.6457)
8 0.0225 (0.0000-0.6446)
9 0.7344 (0.5771-0.8916)
10 0.0000 (0.0000-0.6451)
11 0.0000 (0.0000-0.7910)
12 0.0000 (0.0000-0.6451)