Abstract


Speaker: Yongliang (Vincent) Zhai


Title: Inferring the history of human populations


Abstract: The variations of allele frequencies across populations provide an invaluable window into the history of populations. Single Nucleotide Polymorphism (SNP) frequencies in different populations have been used extensively in the past decade to probe the patterns of expansion of humans across the world via estimation of trees tipped at modern populations. We introduce a new likelihood-based inference method to estimate evolutionary distances for any two populations to their most recent common ancestral population using SNP frequencies. Our model can be applied to both closely related populations and relatively distant populations and this inference method can be easily generalized to a large number of populations with reasonable computation cost. A new type of asymmetric dissimilarity matrix for n populations can be obtained using our method. We propose new algorithms to construct a rooted bifurcating non-clock tree based purely on the asymmetric dissimilarity matrix without extra information on frequencies of ancestral populations. Using our method, we obtained a rooted phylogenetic tree for the 53 human populations of the Human Genome Diversity Panel. Our results are consistent with the prevalent recent single-origin model.