population expansion - senecio7の日記

Keinan A, Mullikin JC, Patterson N & Reich D 2007 Measurement of the human allele frequency spectrum demonstrates greater genetic drift in East Asians than in Europeans. Nat Genet 39:1251-1255.

we explored a model with two bottlenecks and found that it fit both the European and East Asian data sets significantly better
the ancient bottleneck did not have any significant difference in time or intensity between the two populations
the recent bottleneck took place 18 ± 3 kya in Europeans and 16 ± 2 kya in East Asians
we estimated that the populations diverged 17 ± 3 kya, suggesting that the divergence and bottlenecks may have been associated with the same demographic upheavals, perhaps the last glacial maximum
instead of sudden bottleneck, non-African populations might have experienced a long, drawn-out period of genetic drift, comprised of many mild bottlenecks
further analyses will be necessary to distinguish among these hypotheses
we also did not explore the possibility of gene flow between East Asian and North European ancestors after their initial population separation, which might result in an underestimation of the population divergence time
our two-bottleneck model implies a time of ~ 140–80 kya for the first bottleneck (supplementary note), older than the conventionally estimated dates of the out-of-Africa expansion
our time estimates may reflect an averaging of an out-of-Africa bottleneck with earlier events
the 'out-of-Africa' bottleneck may have coincided with the migrations of anatomically modern humans to the Middle East 135–90 kya^20,23-25 rather than with the subsequent European and Asian dispersal
there are also other demographic events in human history that our models are not capturing
for example, the explosive population expansion that occurred in the last ten thousand years

supplementary methods
bottlenecks were modeled as a crash in population size for a fixed number of generations followed by re-expansion to the same effective population size as before the bottleneck, with two parameters capturing the time and inbreeding coefficient of the bottleneck
the inbreeding coefficient, defined here as F = T / 2N, is approximately the probability that two alleles randomly picked from the population after the bottleneck derive from the same ancestral allele just before the bottleneck
the effect of a bottleneck on the frequency spectrum depends primarily on this ratio and is practically independent of the predefined value of T, the number of generation used to model the bottleneck
simultaneous scaling of both T and N does not change the results

supplementary note 1
the modeling estimates no recent expansion
models of expansion do not fit the data significantly better than models of no expansion
a larger sample size should detect the more recent expansions

supplementary note 6
the effect of a bottleneck on the frequency spectrum depends primarily on the inbreeding coefficient, the ratio of its length to (twice) the effective population size during the bottleneck
we modeled a bottleneck using two parameters, the time of the bottleneck and its inbreeding coefficient
we implemented this by keeping the number of generations of the bottleneck fixed on a predefined value of T=100 and varying the effective population size to match the desired inbreeding coefficient values
this note tests to what extent the length of the bottleneck and the effective population size individually, rather than their ratio, affect the frequency spectrum and as a consequence our results
we reanalyzed our data with different values for the number of generations of the bottleneck, scaling the effective population size as the inverse of the number of generations to keep the inbreeding coefficient constant
table 1 presents the maximum likelihood estimates for the bottleneck model for T=10 and T=50
the results are not sensitive to the choice of T