A General Framework for Branch Length Estimation in Ancestral Recombination Graphs.

Area of Science:

• Population Genetics
• Genomic Variation Analysis
• Bioinformatics

Background:

• Ancestral Recombination Graphs (ARGs) are crucial for understanding genomic variation.
• Estimating coalescence times in ARGs often relies on informative priors from coalescent theory, which can introduce bias and complicate subsequent analyses.
• Existing methods may limit the accuracy of downstream inferences.

Purpose of the Study:

• To introduce POLEGON, a novel method for estimating branch lengths in ARGs using an uninformative prior.
• To demonstrate that POLEGON provides more accurate estimates of coalescence times compared to traditional methods.
• To show improvements in downstream inferences, including effective population sizes and mutation rates.

Main Methods:

• Development of POLEGON, a new computational approach for ARG branch length estimation.
• Utilizing an uninformative prior to avoid biases associated with coalescent theory-derived priors.
• Extensive simulations to validate the method's performance across various demographic models.

Main Results:

• POLEGON yields improved estimates of coalescence times in ARGs.
• The method leads to more accurate inferences of effective population sizes under diverse demographic scenarios.
• Downstream analyses, such as mutation rate estimation, are also enhanced.
• Application to 1000 Genomes Project data revealed population-specific histories and mutation signatures.
• Coalescence times exceeding 30 million years were estimated in multiple HLA region segments.

Conclusions:

• POLEGON offers a robust and accurate alternative for estimating ARG branch lengths.
• The uninformative prior approach mitigates bias and improves the reliability of genetic inference.
• This method has significant implications for understanding population genetics, evolutionary history, and disease association studies.