A vectorized method of importance sampling with applications to models of mutation and migration

Area of Science:

• Population Genetics
• Computational Biology
• Statistical Genetics

Background:

• Population genetic data analysis relies on accurate models of population history and evolutionary processes.
• Estimating the likelihood of observed genetic data under complex demographic scenarios is computationally challenging.

Purpose of the Study:

• To develop an efficient importance-sampling method for computing the likelihood of population genetic data configurations.
• To provide a flexible framework applicable to various population history models and state transition processes.
• To demonstrate the utility of the method in analyzing allele frequencies and detecting geographic dispersal.

Main Methods:

• An importance-sampling approach is presented to compute likelihoods for population genetic data.
• The method models state transitions using a Markov chain and assumes independence between coalescence times and genetic states.
• It is applied to gene genealogies from neutral coalescent and linear birth-death processes.

Main Results:

• The method was applied to microsatellite loci in a Sardinian population, rejecting the one-step mutation model for one locus.
• Analysis of low-frequency alleles in the glaucous-winged gull identified geographically restricted dispersal patterns.

Conclusions:

• The developed importance-sampling method offers a robust tool for population genetic inference.
• It facilitates the testing of evolutionary models and the detection of fine-scale population structure and dispersal.
• This approach advances the computational capabilities in understanding population genetic dynamics.