We used two reference panels: (1) the public 1000 Genomes Project Phase 3 haplotypes, version 20130502, (1KGP) (Sudmant et al. 2015 (link)); and (2) The EPIGEN-5M+1KGP reference panel, which is the merge of the 1KGP panel and our unpublished EPIGEN-5M panel, bearing 14,970 more SNPs than the public panel solely. The EPIGEN-5M data set was genotyped with the Illumina HumanOmni5-4v1 array. After quality control, the data set comprises 4,102,271 SNPs for 265 Brazilians from the three cohorts (90, 88, and 87 individuals from Salvador, Bambuí, and Pelotas, respectively) (Supplemental Table S2). We used SHAPEIT2 (Delaneau et al. 2013 (link)) to infer the chromosome phase of the EPIGEN-5M data set (Supplemental Tables S4–S8).