The effect of genotype calling for low or medium coverage data using Tajima’s D . The difference between estimated and known Tajima’s D statistic for three different scenarios with 10 different p-value cutoffs. Each box is estimated on the basis of 100 1 MB regions. Subfigure ‘a)’ is based on genotypes called using the frequency as prior, GC-hwe, and Subfigure ‘b)’ is based on genotypes called using a maximum likelihood approach, GC-mLike. We have included the EB method for the 3 different scenarios in the right side of the figures. Notice that no single best cutoff can be chosen across the three different scenarios for the genotype calling based methods.