108 Results
for author "Jody Hey"
- Table 1Classification of loci under the neutral model with a single group of loci
- Table 2Classification of loci under the true models with two groups of loci
- Table 3Classification of loci under the IM model with selection against gene flow
Figure 1(A–H) Distribution of the posterior modes for the demographic parameters of each scenario considered, under the “migration only” (MO) model. The simulated scenarios are coded as follows: (1) S0, S1, S3, and S5 correspond to cases with zero, one, three, and five loci in group g2 (m = 0.0), respectively; and (2) m values (i.e., m0.05, m0.2, m0.5, and m1.0) are the migration rates for loci in group g1. Horizontal dashed lines correspond to the true parameter values used to simulate the data. For scenarios S1, S3, and S5 the loci in group g2 were simulated with a migration rate of zero to mimic the effects of strong selection against gene flow.
Figure 2Marginal posterior distribution of demographic parameters under the MO model. Densities were obtained by summing the posteriors of the runs where the mean assignment corresponded to the correct classification of loci, under three scenarios with neutral migration rate of m = 0.5 (2NM = 2.5). (A–D) Five selected loci with 2Ns = 50; (E–H) five selected loci with 2Ns = 100; (I–L) five loci in group g2 simulated with m = 0.0 (2Ns ≈ ∞).
Figure 3Posterior distribution of migration rates as a function of the selective coefficients. (A) Mean of the modes of the posterior for migration rates of loci in group g2 (in log scale), defined as the group with lower rates, as a function of the selected coefficients for scenarios with one, three, and five selected loci. (B) Mean difference between the posterior modes of migration rates of groups g1 and g2, as a function of the selective coefficients for scenarios with one, three, and five selected loci.
Figure 4Marginal posterior distribution of demographic parameters obtained with the MO model for runs where the estimated mean assignment was incorrect, i.e., loci were not classified into the correct groups. Densities were obtained by summing the posteriors under the case with neutral migration rate of m = 0.05 (2NM = 0.25). (A–D) Five selected loci with 2Ns = 10; (E–H) five selected loci with 2Ns = 100; (I–L) five loci in group g2 simulated with m = 0.0 (2Ns ≈ ∞).
Figure 5Fit of the estimated likelihood-ratio test statisticto the expected χ2-distributions (solid line). Shown are the ratio of likelihoods integrating over assignment, fF(X|Θ)/fN(X|Θ) (“marginal a”), and the ratio of likelihoods conditional on a fixed random assignment, fF(X|Θ, a)/fN(X|Θ, a) (“conditional a”). (A) MO model with low migration rate m = 0.05. Shown is comparison of the full model with four migration rate parameters, with the nested model with two parameters where migration rates are equal in both groups; i.e.,