Reconstructing Native American migrations using genome-wide sequence data

GRAVEL S; TAINO GENOME SEQUENCING GROUP; MUZZIO M

Cold Spring Harbor

Congreso; Biology of Genomes; 2013

(Muzzio M es parte del Taino Genome Sequencing Group. SIGEVA no permite guardar si no figura en campo individual) We provide the first comprehensive population genetics analysis of the 1000 Genomes populations of Colombian (CLM), Mexican (MXL), and Puerto Rican (PUR) ancestry. Whereas numerous recent studies have focused on reconstructing Native American population histories, these were limited to SNP chip and Y-chromosome data. The 1000 Genomes whole-genome sequencing data offers an opportunity to refine demographic inferences including population sizes and split times. To do this, we introduce several methods to disentangle the genomic contributions of African, European, and Native American ancestry in these populations, using allele frequencies, IBD, and local ancestry estimations. Inferred Native American range from 13% in PUR to over 50% in MXL. Focusing on genomic segments assigned to have Native Ancestry, we find that Mexican Native ancestry clusters with Northern and Central Amerind Native groups, and that the CLM Native Ancestry clusters with Chibchan-Paezan-speaking populations. The PUR ancestry appears closer to populations from Equatorial-Tucanoan populations, supporting a South American origin of the Native ancestry in Puerto Ricans, supporting archeological records. Based on a new methodology, we estimate the allele frequencies in the Native American fraction of the populations, and model their distribution using a three-population demographic model. This reveals that the ancestral populations to the three groups likely split in close succession a few hundred years apart: the most likely scenario, based on a peopling of the Americas 15 kya, supports that the MXL ancestors split 11.5kya, with a subsequent split of the ancestors to CLM and PUR 11.0 kya. The model also features very different Native populations sizes, with a Mexican population of 70,000, a Colombian population of 8000, and a Puerto Rican population of 2000, indicating a strong subsequent bottlenecks during the colonization of South America. Supporting this ancient split, we find very little cross-population identity-by-descent within blocks of Native American ancestry: the 100 longest cross-population IBD tracts are within European-assigned tracts, reflecting the higher relatedness of European founders in the Americas, and possibly their greater mobility. Our modeling demonstrates that whole-genome sequence from populations with complex histories can be leveraged to understand the history of each ancestral population in much greater detail. This is particularly useful for Native populations from Puerto Rico, whose descendants carry extensive genetic contributions from Europe and Africa. Taken together with the Monte Verde archeological site, our results support multiple early splits during the rapid settlement of the Americas.