A Comparative Analysis of Genetic Ancestry and Admixture in the Colombian Populations of Choco and Medellin

At least 20% of Colombians identify as having African ancestry, yielding the second largest population of Afro-descendants in Latin America. To date, there have been relatively few studies focused on the genetic ancestry of Afro-Latino populations. We report a comparative analysis of the genetic ancestry of Choco, a state located on Colombia's Pacific coast with a population that is >80% Afro-Colombian. We compared genome-wide patterns of genetic ancestry and admixture for Choco to six other admixed American populations, with an emphasis on a Mestizo population from the nearby Colombian city of Medellin. One hundred sample donors from Choco were genotyped across 610,545 genomic sites and compared with 94 publicly available whole genome sequences from Medellin. At the continental level, Choco shows mostly African genetic ancestry (76%) with a nearly even split between European (13%) and Native American (11%) fractions, whereas Medellin has primarily European ancestry (75%), followed by Native American (18%) and African (7%). Sample donors from Choco self-identify as having more African ancestry, and conversely less European and Native American ancestry, than can be genetically inferred, as opposed to what we previously found for Medellin, where individuals tend to overestimate levels of European ancestry. We developed a novel approach for subcontinental ancestry assignment, which allowed us to characterize subcontinental source populations for each of the three distinct continental ancestry fractions separately. Despite the clear differences between Choco and Medellin at the level of continental ancestry, the two populations show overall patterns of subcontinental ancestry that are highly similar. Their African subcontinental ancestries are only slightly different, with Choco showing more exclusive shared ancestry with the modern Yoruba (Nigerian) population, and Medellin having relatively more shared ancestry with West African populations in Sierra Leone and Gambia. Both populations show very similar Spanish ancestry within Europe and virtually identical patterns of Native American ancestry, with main contributions from the Embera and Waunana tribes. When the three subcontinental ancestry components are considered jointly, the populations of Choco and Medellin are shown to be most closely related, to the exclusion of the other admixed American populations that we analyzed. We consider the implications of the existence of shared subcontinental ancestries for Colombian populations that appear, at first glance, to be clearly distinct with respect to competing notions of national identity that emphasize ethnic mixing (mestizaje) vs. group-specific identities (multiculturalism).