PMC

Bayesian skyline plots of population size change through time for regional groups. Two curves are shown for the NRY data, based on ‘fast’ and ‘slow’ mutation rate estimates.

Distribution of N_f and N_m values, based on simulations. The density of the top 1% of the posterior values obtained from simulations of the mtDNA and NRY sequences are shown. (A) ancestral effective population sizes; (B) current effective population sizes. The dashed line in each plot follows a 1:1 ratio.

Diversity and AMOVA results. (A) Mean number of pairwise differences (and SE bars) for the NRY and mtDNA sequences from each regional group. (B) AMOVA results for the entire worldwide dataset, and for each regional group of populations. Two comparisons are shown for the entire dataset; the left comparison includes regional groups as an additional hierarchical level, while the right one does not. * indicates that the among-population component of diversity does not differ significantly from zero (after Bonferroni adjustment of the P value for multiple comparisons).

Bayesian trees and divergence time estimates for mtDNA and NRY haplogroups. (A) mtDNA haplogroups; (B) NRY haplogroups with the fast mutation rate; (C) NRY haplogroups with the slow mutation rate. Red asterisks denote nodes with low support values (<0.95). F* in the NRY trees indicates a sample that was assigned to haplogroup F by SNP genotyping, but does not fall with other haplogroup F samples. Some NRY haplogroup K samples formed a monophyletic clade (labelled K in the trees) while others fell with haplogroup M samples (labelled KM in the trees); see also Additional file : Figure S8.

Pictorial representation of the divergence time and female and male effective population size estimates, based on the simulation results. Red numbers reflect N_f (with ancestral N_f at the point of the red triangle and current N_f at the base of the red triangle) and blue numbers correspondingly reflect ancestral and current N_m. The numbers in the black oval indicate the founding effective sizes for the initial out-of-Africa migration, and dates on arrows indicate divergence times based on the model in Figure . Arrows are meant to indicate the schematic direction of migrations and should not be taken as indicating literal migration pathways, for example, the results indicate divergence of the ancestors of Oceanians 61,000 years ago, but not the route(s) people took to get to Oceania.

The model of population history used in simulations. We assumed a single out-of-Africa migration and further population divergence events (see text for further details). The model begins with the ancestral population in Africa (at time T1), a single out-of-Africa migration (T2), the first split between Oceania and Eurasia (T3), then Europe and Asia (T4), followed by Central and East Asia (T5), and finally between East Asia and the Americas (T6). We also required T2 to be greater than T3. The model assumes no migration between regions following divergence; in support of this assumption, there is very little sequence sharing between regions. We do allow changes in population size. This model was first used to estimate divergence times with combined mtDNA and NRY sequences, then the model and estimated mean divergence times were used in separate simulations of the mtDNA and NRY sequences to estimate ancestral and current N_f and N_m.