TABLE 4

An evaluation of alternative demographic models for the total population and the Baltico–Nordic and the Alpine domains

TotaleBaltico–Nordic domaineAlpine domaine
Mean πaMean DMean HMean πMean DMean HMean πMean DMean H
Observed1.47−0.88 [0.38]−0.74 [3.84]1.53−0.55 [1.00]−0.27 [3.28]1.59−0.66 [0.45]−0.26 [1.45]
Model
    SNMb1.47−0.02 (<10−4)−0.000 (0.015)1.53−0.03 (0.005)−0.00 (0.180)1.59−0.03 (0.001)−0.00 (0.209)
    Growthc1.47−0.80 (0.337)0.68 (<10−4)1.54−0.82 (0.955)0.73 (0.000)1.59−0.80 (0.815)0.76 (0.000)
    Bottleneckd1.45−0.28 (0.079)−1.75 (0.693)1.53−0.31 (0.304)−1.81 (0.901)1.59−0.29 (0.233)−1.96 (0.860)
  • P-values for the observed means under the model simulated are given in parentheses. The numbers within brackets are the variances across loci of the parameters.

  • a Average π per locus across loci (the average number of sites surveyed is 719 bp), mean Tajima's D across loci, and mean Fay and Wu's H across loci.

  • b Standard neutral model.

  • c The growth rate was G = 10. θ = 4.78.

  • d We assumed a population shrinking at rate 10 up to time t1 = 0.003 × 4Ne before present (this represents population growth in the forward direction), then going through a bottleneck of severity f = 0.0005 until t2 = 0.0035 × 4Ne, and then having an ancestral population the same size as the current population. Assuming that Ne = 500,000 and a generation time of 25 years, t1 = 150,000. If we assume Ne = 106, t1 = 300,000. In the first scenario the bottleneck would last 25,000 years. θ = 10.03.

  • e The analysis was based on 21 loci in the total data set and the Baltico–Nordic domain and 19 loci in the Alpine domain as phynrII and phyP2 were monomorphic in the latter.