Probability distribution of haplotype frequencies under the two-locus Wright-Fisher model by diffusion approximation

Area of Science:

• Population Genetics
• Computational Biology
• Evolutionary Biology

Background:

• Haplotype frequency distribution is fundamental to population genetics.
• Calculating these distributions under forces like selection and drift is computationally challenging for large populations using classical models.
• The Wright-Fisher model can be approximated by diffusion processes, but exact solutions for Kolmogorov equations are elusive.

Purpose of the Study:

• To develop an accurate and efficient numerical method for computing haplotype frequency distributions.
• To address the computational intractability of exact solutions for Kolmogorov equations in diffusion approximations.
• To provide a tool applicable to transient states and models incorporating selection or mutation.

Main Methods:

• Developed a numerical method based on finite differences to solve Kolmogorov equations.
• Applied the method to approximate the Wright-Fisher diffusion process.
• Validated the method's accuracy for computing conditional joint densities of haplotype frequencies.

Main Results:

• The finite difference method accurately computes conditional joint haplotype frequency densities.
• The method is significantly faster than Monte Carlo simulations.
• It is applicable to complex scenarios including selection and mutation.

Conclusions:

• The developed numerical method offers an accurate and computationally efficient solution for population genetics problems.
• This method advances the study of haplotype frequency dynamics under various evolutionary forces.
• It provides a valuable alternative to computationally intensive simulation techniques.