Leveraging whole genome sequencing data for demographic inference with approximate Bayesian computation.

Area of Science:

• Evolutionary Biology
• Population Genetics
• Bioinformatics

Background:

• Accurate inference of historical demographic features like gene flow and divergence times is crucial in evolutionary biology.
• Approximate Bayesian computation (ABC) is a common method for estimating demographic parameters, but often relies on limited genomic data (e.g., microsatellites, RADseq).
• The utility and optimal implementation of whole genome sequencing (WGS) data within ABC frameworks remain areas of active research.

Purpose of the Study:

• To evaluate the potential improvements in demographic parameter estimation accuracy using WGS data in ABC analyses.
• To quantify the impact of increased data volume, haplotype-based summary statistics, and locus length on estimation accuracy.
• To assess the influence of incorporating locus-specific recombination rates and background selection information into ABC models.

Main Methods:

• Utilized pseudo-observed data sets to simulate and test different genomic data scenarios.
• Compared parameter estimation accuracy between a hypothetical RADseq dataset (2.5 Mbp) and a WGS dataset (1 Gbp with 100 Kbp sequences).
• Investigated the effects of including locus-specific recombination rates and background selection on ABC model performance.

Main Results:

• A 1 Gbp WGS dataset, particularly with haplotype statistics and increased data, led to substantial gains in parameter estimation accuracy compared to a 2.5 Mbp RADseq dataset.
• Haplotype-based summary statistics and increased data volume were primary drivers of improved accuracy.
• Ignoring background selection or assuming uniform recombination negatively impacted accuracy in numerous cases.

Conclusions:

• Whole genome sequencing data offers significant advantages for demographic inference using Approximate Bayesian computation.
• Incorporating haplotype information and considering locus-specific recombination and background selection are critical for accurate demographic modeling.
• The findings provide a methodological validation for future demographic history analyses utilizing large-scale genomic data.