Admixture f3 statistics based on ancient and present-day genomes

We inferred MJS’s genetic lineage using admixture f3-statistics, a method based on measuring the allele frequency correlations between populations (Patterson et al. 2012). To maximize the comprehensiveness of this analysis and evaluate genetic associations, we selected 2670 present-day and 108 ancient genomes (using published ancient genomes (Lazaridis et al. 2014; Lipson et al. 2018; Keller et al. 2012; Jones et al. 2015; Siska et al. 2017; Haak et al. 2015; Allentoft et al. 2015; Skoglund et al. 2014; Fu et al. 2014; Raghavan et al. 2014; Seguin-Orlando et al. 2014; Gamba et al. 2014)). To measure genetic associations, we used a notation f3(A,B;Kazakh), where A and B were ancient and present-day populations in various combinations. We employed the qp3PopTest program (v300) from the ADMIXTOOLS (v3) package to calculate the f3 statistics. We extracted the combinations which had high significance (|Z| > 3) and a sufficient number (> 1000) of examined SNPs (Table S2). The top 30 genome pairs with the most negative f3 values are plotted (“Results and discussion” section).