Estimating divergence parameters with small samples from a large number of loci

Area of Science:

• Population Genetics
• Genomics
• Evolutionary Biology

Background:

• Traditional methods for studying divergence with gene flow often use many individuals at few loci, limiting insights into ancient events.
• The increasing availability of whole-genome sequence data offers a new approach, utilizing numerous loci from fewer individuals for studying ancient divergence.

Purpose of the Study:

• To develop a novel maximum-likelihood method for genomic data analysis under the isolation-with-migration model.
• To enable precise likelihood calculations through numerical integration, avoiding Monte Carlo sampling of genealogies.
• To assess the impact of mutation rate variation on divergence estimates.

Main Methods:

• Developed a new maximum-likelihood method tailored for large genomic datasets (many loci, few individuals).
• Employed numerical integration over all genealogies for precise likelihood calculation, contrasting with Monte Carlo methods.
• Evaluated two models for mutation rate variation, identifying superior performance with random variable treatment.

Main Results:

• The new method demonstrated effectiveness on simulated genomic datasets.
• Treating mutation rates as random variables improved the accuracy of divergence estimates.
• Applied to Drosophila melanogaster and D. simulans, the method revealed a significant, albeit low, gene flow signal from D. simulans to D. melanogaster.

Conclusions:

• The developed maximum-likelihood method is effective for analyzing genomic data to study ancient divergence and gene flow.
• Accounting for mutation rate variation is crucial for accurate evolutionary inference.
• Evidence of historical gene flow between D. melanogaster and D. simulans was detected, contributing to our understanding of their evolutionary history.