Unraveling subcontinental ancestries across Africa, Americas and Europe, and implications for admixture/association mapping of complex traits

Morehouse School Of Medicine, Atlanta GA

Investigators

Abstract

Abstract Genomic ancestries are known to be associated with different biological traits, which is partially due to ancestral differences in genetic susceptibility overrepresented among certain ancestries. The lack of knowledge of population genomics on subcontinental scales leads to the simplistic assumption that continental ancestries (e.g., African, European, and Native American ancestries) are single homogeneous entities, rather than a mix of subcontinental ancestries. As a result, populations recognized as admixed are often excluded from genomic studies due to concerns over population stratification, and populations not recognized as admixed are analyzed using sub-optimum statistical genomic models. Potential consequences are that detection of causal genetic variation is hampered and estimation of effect sizes is biased. There are two major approaches underlying ancestry inferences: Global individual ancestry, which is the estimate of individuals’ ancestry proportions based on their genomes; and local ancestry (LA), which is ancestry estimates of individuals’ chromosomal segments, and it is based on the fact that an admixed individual's genome is a mosaic of segments from different ancestral origins. Admixture mapping (AM) is a powerful gene mapping technique based on LA for identifying loci that confer differential risk by ancestry in ancestrally diverse populations such as continental Africans, African Americans, and Latinos. However, current AM approaches only consider continental population origins (e.g., African vs. European ancestry in broadly defined terms), and consequently, AM approach is being almost exclusively applied to continental admixed populations (e.g. African Americans and Latinos), despite most human populations being mixed at a subcontinental level. In Aim 1 of this award, I will compile the largest genome-wide data collection (~75,000 ancestrally diverse individuals) and apply state-of-the-art methods in population genomics to maximize the power to detect and define subcontinental ancestries across Africa, the Americas, and Europe. Further, I will investigate how subcontinental African, European and Native-American ancestries have admixed to produce the current genetic profile of admixed populations in the Americas (Aim 2). In Aim 3, I will utilize subcontinental local ancestry inference to improve AM power to discover genetic associations. By completing each one of my proposed specific aims, this grant will provide several innovations, as follows: 1) to reveal a fine-scale landscape of subcontinental ancestries across three continents: Africa, America, Europe; 2) to gain a comprehensive understanding of admixture dynamics in the Americas on a subcontinental scale (Aim 2); and 3) to extend admixture mapping (AM) beyond the traditional approach, by making AM applicable to both admixed and “non-admixed” individuals. Importantly, the use of subcontinental ancestries will make admixture mapping applicable to its fullest extent, while also allowing for full control (global and local ancestry levels) of confounding by population structure for the first time in the 15 years of the GWAS era. These large-scale data analyses and advanced training acquired during this grant will provide me the necessary practical and didactical experience to establish a robust independent research program at the interface of human population genetics, bioinformatics, and genetic epidemiology.