Figure 2.
—The effects of population properties on the mean and RMSE of admixture proportion estimates (*p̂*_{1}) from different estimators (indicated by different lines shown at the top right corner of the graph). Both *x*- and *y*-axes are in logarithm scale. All the simulation results are obtained using parameters *n*_{i} = *N*_{j} = 5000 (*i* = 0, 1, 2; *j* = 1, 2, h), *S*_{j} = 50, and *L* = 1 for a DNA sequence. (A and B) The effects of divergence time (*t*_{i}, t_{1} = *t*_{2}) of parental populations, with results obtained from 5000 (for *t*_{i} < 0.1), 2000 (for 0.1 ≤ *t*_{i} < 1), or 1000 (for *t*_{i} ≥ 1) replicates using various values of ξ and parameters ψ= 100, *p*_{1} = 0.2, and *U* = 0.0001. (C and D) The effects of genetic drift after the admixture event (*T*_{j}, T_{1} = *T*_{2} = *T*_{h}) on the mean and RMSE of *p̂*_{1}, with results obtained from 500 (for *T*_{j} < 0.01), 1000 (for 0.01 ≤ *T*_{j} < 0.1), or 5000 (for *T*_{j} ≥ 0.1) replicates using various values of ψ and parameters ξ= 5000, *p*_{1} = 0.2, and *U* = 0.0001. (E and F) The effects of the true admixture proportion (*p*_{1}) on the mean and RMSE of *p̂*_{1,} with results obtained from 1000 replicates using various values of *p*_{1} and the parameters ξ= 5000, ψ= 100, and *U* = 0.0001. (G and H) The effects of the mutation rate (θ = 4*NU*) of a DNA sequence on the mean and RMSE of *p̂*_{1}, with results obtained from 5000 (for θ< 1), 500 (for 1 ≤θ< 10), or 1000 (for θ≥ 10) replicates using various values of θ and the parameters ξ= 5000, ψ= 100, and *p*_{1} = 0.2.