Estimation of Natural Selection and Allele Age from Time Series Allele Frequency Data Using a Novel Likelihood-Based...

Area of Science:

• Population genetics
• Evolutionary biology
• Computational biology

Background:

• Temporally spaced genetic data are crucial for understanding population genetic parameters and natural selection.
• Existing methods often assume alleles arise at low frequencies, which may not always be accurate.

Purpose of the Study:

• To develop a novel likelihood-based method for jointly estimating selection coefficient and allele age from time series allele frequency data.
• To overcome limitations of existing methods by not requiring assumptions about allele origin frequency.

Main Methods:

• Utilized a hidden Markov model (HMM) based on Wright-Fisher diffusion, conditioned on allele survival.
• Calculated likelihood by numerically solving the Kolmogorov backward equation and reweighting with emission probabilities.
• Reduced the estimation problem to a one-dimensional search for the selection coefficient.

Main Results:

• The method accurately estimates selection coefficients and allele ages under various demographic histories through simulations.
• Application to ancient horse DNA data demonstrated the method's utility.
• Ignoring demographic histories or sample grouping can lead to biased inference.

Conclusions:

• The developed method provides accurate joint estimation of selection coefficient and allele age from time series genetic data.
• This approach is robust to demographic fluctuations and offers an improvement over existing techniques.
• Accurate demographic modeling is essential for reliable inference from genetic data, particularly ancient DNA.