Correcting coalescent analyses for panel-based SNP ascertainment

Area of Science:

• Population Genetics
• Genomic Data Analysis
• Bioinformatics

Background:

• Single-nucleotide polymorphism (SNP) data are commonly generated using SNP chips derived from small initial sequencing panels.
• Chip design often involves removing low-frequency alleles, potentially leading to information loss and biased downstream analyses.
• Accurate estimation of population genetic parameters, such as scaled population size (Θ), is crucial for understanding evolutionary processes.

Purpose of the Study:

• To quantify the information loss associated with SNP panel ascertainment and rare allele omission in coalescent analyses.
• To develop and evaluate correction methods for coalescent estimation of population size (Θ) using SNP data from panels.
• To assess the impact of recombination and migration on Θ estimates in multi-population data.

Main Methods:

• Utilized coalescent estimation to model the scaled population size parameter (Θ).
• Developed and applied corrections for panel ascertainment bias and the omission of rare alleles.
• Extended the methodology to analyze recombinant multiple population data, incorporating recombination and migration effects.

Main Results:

• Demonstrated significant information loss due to SNP panel ascertainment and rare allele removal.
• Showed that incorporating known panel size improves Θ estimates but increases computational cost.
• Found that corrected estimates for panel ascertainment and rare allele omission largely correct biases, but remain less accurate than estimates from fully sequenced data.

Conclusions:

• Coalescent analysis of SNP data from panels requires specific corrections to mitigate ascertainment biases.
• While corrections improve accuracy, fully sequenced data provides more precise estimates of population genetic parameters.
• The developed methods are applicable to complex population structures, accounting for recombination and migration.