We then moved on testing the conclusions drawn from EstBB on the UK Biobank27 (UKBB), particularly focusing on the samples for which at least two major ancestry components could be detected, and for whom phenotypes were available. We used the first six principal components to select 5000 samples from the core of European UKBB individuals27, and to define 34212 samples as genetically non-European and putatively admixed based on their distance from the main unadmixed core (Supplementary Fig. 4a). We further explored the ancestry makeup of these individuals through an ADMIXTURE32 analysis, projecting the individuals onto the allele frequency spectrum inferred by the software for a set of 1000 Genomes samples, for which global ancestry and country of origin is known (Supplementary Fig. 4b). By this way we confirmed the admixed/unadmixed nature of the non-European samples selected from the UKBB and we further labeled the observed major ancestry components. We divided the selected UKBB samples into genetically “African”, “East Asian”, “European” and “Admixed” individuals, based on their PCA and ADMIXTURE scores, and used the former three groups as sources to perform ancestry deconvolution on the latter, see Supplementary Table 2 for sample sizes.
