Skip to main content
. 2012 Nov 28;110(4):363–371. doi: 10.1038/hdy.2012.102

Table 2. Mean and s.d. of four summary statistics calculated using neutral variants at three time points during the bottleneck.

    π ( × 10−3)
θW ( × 10−3)
Tajima's D
Fu and Li's D
τ Region Forward Coalescent Forward Coalescent Forward Coalescent Forward Coalescent
<0 Flanking 8.4 (2.1) 8.6 (2.1) 8.5 (1.6) 8.7 (1.7) −0.066 (0.50) −0.092 (0.49) −0.106 (0.61) −0.143 (0.61)
  Embedded 8.1 (3.2) 8.2 (3.3) 8.2 (2.6) 8.4 (2.7) −0.095 (0.72) −0.115 (0.73) −0.125 (0.85) −0.154 (0.87)
0.02 Flanking 7.9 (2.1) 8.1 (2.1) 7.4 (1.6) 7.6 (1.7) 0.272 (0.55) 0.227 (0.55) 0.467 (0.64) 0.402 (0.66)
  Embedded 7.7 (3.2) 7.8 (3.3) 7.2 (2.5) 7.4 (2.6) 0.216 (0.78) 0.181 (0.77) 0.339 (0.85) 0.287 (0.85)
0.325 Flanking 9.2 (2.1) 9.3 (2.1) 10.3 (1.7) 10.5 (1.7) −0.457 (0.44) −0.470 (0.43) −0.708 (0.59) −0.739 (0.59)
  Embedded 8.8 (3.2) 9.0 (3.3) 9.9 (2.7) 10.1 (2.8) −0.449 (0.65) −0.478 (0.65) −0.621 (0.85) −0.663 (0.85)

Standard deviations are shown in parentheses. The model used to produce Figure 4 was simulated. Random samples of size 20 were generated at three time points (indicated with black, red and blue arrows in Figure 4a). The statistics were calculated using sequence variability at two types of neutral sites—those in the two flanking regions (Flanking) and those embedded in the centrally located selected region (Embedded).