An approximate likelihood for genetic data under a model with recombination and population splitting.

Area of Science:

• Population Genetics
• Statistical Genetics
• Computational Biology

Background:

• Accurate modeling of population splits is crucial for understanding evolutionary history.
• Existing models struggle with analyzing loosely-linked haplotype data and non-equilibrium population structures.
• The 'Product of Approximate Conditionals' and 'copying model' provide a foundation for new statistical approaches.

Purpose of the Study:

• To introduce a novel approximate likelihood for population genetic data.
• To develop a method capable of analyzing loosely-linked haplotype data with non-equilibrium population structure.
• To leverage both allele frequency differences and haplotype sharing for inferring population split timing.

Main Methods:

• Developed a new approximate likelihood based on the 'Product of Approximate Conditionals' and a 'copying model'.
• Incorporated the effects of recombination into the copying model.
• Applied the method to analyze loosely-linked haplotype data, extracting information from shared haplotype lengths.

Main Results:

• The new method allows for efficient approximate likelihood-based analyses of unlinked and loosely-linked data.
• It effectively utilizes information from both allele frequency differences and shared haplotype lengths.
• The approach addresses limitations of previous models for non-equilibrium population structures.

Conclusions:

• The developed method offers an efficient tool for population genetic analyses, particularly for loosely-linked haplotype data.
• It enhances the ability to infer population split timing by incorporating haplotype structure.
• Future extensions could include identifying migrant haplotypes, further refining population history inferences.