Fig. 2: Population structure.
From: Indigenous Australian genomes show deep structure and rich novel variation
a, Location and sample size for all Australian and Papuan samples. b, Hierarchical clustering of unrelated individuals on the basis of pairwise outgroup F3 statistic values. Colour corresponds to sampling location. c, ADMIXTURE-inferred ancestry for unrelated individuals allowing seven clusters, ordered according to sampling location. Colour was assigned to each cluster post hoc on the basis of the scheme in a and the majority membership of each cluster. d, Pairwise sharing of rare alleles (above diagonal) and IBD (below diagonal) tracts among all individuals. Counts were rescaled according to the proportion of the genome missing due to ancestry masking in each pairwise comparison. Comparisons between first- and second-degree relatives are indicated in red. e, UMAP clustering of unrelated individuals on the basis of minor allele frequency-corrected COV distances, reduced to the first ten components by MDS. Box expands the positions of Tiwi Island individuals. f, Clustering of Tiwi individuals on the basis of co-ancestry values estimated using fineSTRUCTURE run on all unrelated and unadmixed samples (see Extended Data Fig. 4a for the full tree). Light blue (Bathurst Island) and dark blue (Melville Island) indicate sampling location, and yellow and grey indicate cluster membership.
