Evaluating the performance of a multilocus Bayesian method for the estimation of migration rates

Area of Science:

• Molecular Ecology
• Population Genetics
• Bioinformatics

Background:

• Bayesian methods are increasingly used in molecular ecology for demographic parameter estimation.
• Existing method evaluations often lack thoroughness, necessitating further performance analysis.

Purpose of the Study:

• To comprehensively evaluate the performance of the bayesass method for estimating recent migration rates and inbreeding coefficients.
• To identify the parameter space (migration rates, FST) where bayesass provides accurate estimates.

Main Methods:

• Expanded simulation studies using multi-allelic markers and varied population numbers.
• Investigated the impact of varying migration rates and FST on estimation accuracy.
• Assessed the reliability of posterior assignment probabilities for individual assignments.

Main Results:

• Bayesass provides accurate migration rate estimates when model assumptions hold and FST is >= 0.05, even for high migration rates (approx. 0.1).
• When model assumptions are violated, accurate estimates require very low migration rates (m=0.01) and high FST (>= 0.10).
• High posterior assignment probabilities do not guarantee accurate individual assignments.

Conclusions:

• Bayesass is a valuable tool for molecular ecology but requires careful consideration of model assumptions and genetic differentiation.
• Researchers should be cautious when interpreting posterior assignment probabilities due to potential inaccuracies.
• Further validation of population genetics methods is crucial for reliable demographic inference.