3.2 | Simulating the observed data with MetHis
Using MetHis , we produced 90,000 vectors of 24 summary statistics
each, overall highly consistent with the observed ones for the ACB and
the ASW populations respectively. First, each observed statistic is
visually reasonably well simulated under the nine competing scenarios
here considered (Supplementary Figure S7 ). Second, the observed
data each fell into the simulated sets of 24 summary statistics
projected in the first four PCA dimensions (Supplementary Figure
S8 ). Finally, the observed summary statistics vectors were not
significantly different (p-value = 0.468 and 0.710, for the ACB and ASW
respectively) from the simulated ones using a goodness-of-fit approach
(Supplementary Figure S9 ). Therefore, we successfully simulated
datasets producing sets of summary statistics reasonably close to the
observed ones, despite considering constant effective population sizes,
fixed virtual source population genetic pool-sets, and neglecting
mutation during the admixture process.