TABLE 1

The effect of different null distributions on CLRT P-values, applied to the Netherlands data in Beisswanger et al. (2006)

Null modelAscertainmentEmbedded ImageCLRT P-value
SNMaIgnoredNA0.054
SNMAccounted forb0.0250.99
SNMAccounted forc0.024d0.99
BottleneckeIgnored0.998f0.27
BottleneckeAccounted for0.414g0.81
  • a SNM, standard neutral model. When ascertainment of regions is ignored, this is the standard CLRT (Kim and Stephan 2002), and Embedded Image is not relevant.

  • b Calculated using rejection sampling and Equation 10.

  • c Calculated using Equation 11, see text for details.

  • d Calculated as the mean of Embedded Image, from 105 simulated replicates.

  • e Using the parameters from Thornton and Andolfatto (2006), tr = 0.004, d = 0.015, f = 0.03 (see Figure 2), and the mutation and recombination rates used in Beisswanger et al. (2006).

  • f Calculated using Equation 10, with Embedded Image.

  • g Calculated using Equation 10, with Embedded Image.