Figure 5

Results of the PCA, ADMIXTURE, and TreeMix analyses, indicating gene flow from India to Australia. (A) PCA of aboriginal Australians (AUA), highlanders of Papua New Guinea (NGH), Europeans (CEU), and 26 Indian populations. PC1 is driven by differences between the populations of Sahul and Eurasia. PC2 reflects a north-to-south gradient of European ancestry observed in Indian groups, with the southernmost group being the Onge, a Negrito population from the Andaman islands. (B) Population structure estimated using ADMIXTURE for K = 4. Each vertical bar represents an individual, and each color describes the proportion of each individual’s genome that comes from one of the four hypothetical ancestral populations (K). (C) Population graph obtained with TreeMix. First, the maximum likelihood tree of the nine populations included in the analysis was inferred, and then, migration events were added to the tree sequentially, until a graph with the smallest residuals was found. The graph that best fits the data has four inferred migration edges. Populations included are: AUA, NGH, Onge, Mamanwa (a Negrito group from the Philippines; MWA), East Asians (Chinese, CHB), Island Southeast Asians (Borneo, BOR), Indian populations: Chenchu, Kurumba, and Dravidian speakers from South India (DRA). Reprinted with permission from Pugach et al. [68].