All analyses were performed using PLINK v.107 [15]. The number of individuals from each population is
reported in Supplementary Table 2. Simulations were used to assess the power to detect ancient or recent admixture. In all our simulations we used unlinked markers for two reasons: first, the main analyses used were ADMIXTURE [16], the three-population test [17], TREEMIX [18] and Principal Components Analysis (PCA), which all assume unlinked markers; second, the probability to find a segment of x cM (from the source population) λ generations after admixture selleck chemicals is 1 − (1 − e(−λx)), so we estimated that 90% of the fragments remaining after 6.000 years would be shorter than 50 kb, so considering the level of linkage disequilibrium could be considered as single loci. One simulation approach was used to estimate the minimum threshold of recent admixture that would be detectable. We selected 5000 unlinked markers from the JPT and Ecuadorian SNP genotypes, and created artificial genomes with different levels of markers coming from one population. In detail, we simulated 16 admixed Ecuadorians with 50%, 20%, 10%, 5% or 1% JPT admixture; the simulated admixed individuals were then analyzed using ADMIXTURE v.122 with Ecuadorian and Japanese as reference populations. Simulations to evaluate the power to detect selleck inhibitor a single
more ancient admixture event were performed using the simuPOP python library [19], using parameter values for effective population size and populations split times obtained from the SNP genotype data using the procedure learn more of McEvoy [20] implemented in the NeON R package available at http://www.unife.it/dipartimento/biologia-evoluzione/ricerca/evoluzione-e-genetica/software. We modelled a single pulse of migration from a Source population
(representing the East Asian population) to produce an Admixed population (representing the Ecuadorian population); an additional population was simulated as a control (representing an unmixed Native American population). The probability for one individual to migrate from the Source population to the Admixed population was set at 0%, 1%, 5% or 10%. For the 10% scenario, individuals were sampled before the migration event, immediately after the migration event, and at the present time, 6 Ky later. The sample size used was 50 individuals, the genome considered consisted of 2200 independent loci on 22 chromosomes; each scenario was replicated 100 times. Each replicated dataset was analyzed using ADMIXTURE v.122. Principal Components Analysis was carried out using EIGENSOFT v.5.0.2 [21]. Ecuadorian and JPT samples were projected onto the axes obtained from all HGDP populations. PCA was performed on two different datasets: first, with all the populations in this study, and second with just the Native Americans (including the Ecuador samples), Japanese (including JPT), Yakut, French and Russian samples.