Population-genetic inference from pooled-sequencing data.

Area of Science:

• Population Genetics
• Genomic Analysis
• Statistical Bioinformatics

Background:

• Pooled-population sequencing is common for allele frequency estimation.
• Existing methods often lack a rigorous statistical foundation.
• Sequencing and sampling complexities can introduce biases.

Purpose of the Study:

• Develop a statistical framework for allele frequency estimation in pooled sequencing.
• Provide methods for hypothesis testing, including monomorphism and inter-population differences.
• Address the impact of sequencing errors and multi-level sampling.

Main Methods:

• Introduced a closed-form, maximum-likelihood estimator for allele frequencies.
• Developed a likelihood-ratio test for monomorphism.
• Formulated a framework for comparing allele frequencies between populations considering two tiers of sampling.

Main Results:

• Demonstrated that unbiased allele frequency estimates are challenging to achieve.
• Identified a minimum minor allele frequency threshold for reliable polymorphism detection.
• Showed that analyses ignoring sampling tiers lead to high false-positive rates.
• Found that detecting allele frequency differences requires substantial sequencing depth and individual sampling (>100).

Conclusions:

• A robust statistical framework is essential for accurate pooled sequencing analyses.
• Accounting for sequencing errors and hierarchical sampling is critical.
• Underpowered studies (low individual count or sequencing depth) limit the detection of population differences.