WFABC: a Wright-Fisher ABC-based approach for inferring effective population sizes and selection coefficients from...

Area of Science:

• Population Genetics
• Evolutionary Biology
• Bioinformatics

Background:

• Advancements in sequencing technologies provide increasingly available time-sampled population genetic data.
• Inferring population genetic parameters from temporal data is crucial for understanding evolutionary processes.
• Existing methods based on the Wright-Fisher model have limitations in accurately estimating selection coefficients (s) and effective population size (Ne).

Purpose of the Study:

• To compare and analyze four recent Wright-Fisher model-based approaches for inferring selection coefficients from simulated temporal data.
• To demonstrate the advantage of a novel approximate Bayesian computation (ABC)-based method for inferring genomewide average effective population size (Ne).
• To apply the ABC method to infer per-site selection coefficients and investigate allele frequency dynamics in a real-world dataset.

Main Methods:

• Simulation of temporal population genetic data sets.
• Comparison of four Wright-Fisher model-based inference methods for selection coefficients.
• Implementation and application of a novel approximate Bayesian computation (ABC) method for estimating effective population size (Ne) and selection coefficients (s).
• Estimation of the dominance ratio (h) to model allele frequency variation.

Main Results:

• The approximate Bayesian computation (ABC) method accurately infers genomewide average effective population size (Ne) from time-serial data.
• Using Ne as a prior significantly improves the accuracy and precision of per-site selection coefficient (s) inference.
• Application to the Panaxia dominula dataset suggests a recessive lethal model best explains the observed allele frequency variation, with an estimated dominance ratio (h).

Conclusions:

• The developed ABC method offers a powerful tool for robust inference of population genetic parameters from time-series data.
• Accurate estimation of effective population size (Ne) is critical for precise inference of selection coefficients (s).
• The findings provide insights into the evolutionary dynamics of the medionigra genotype in Panaxia dominula, highlighting the importance of genetic models incorporating dominance.