Polarising SNPs Without Outgroup.

Area of Science:

• Population genetics
• Computational biology
• Bioinformatics

Background:

• Allele polarization is crucial for population and quantitative genetics.
• Current methods rely on outgroup data, which can be problematic with distant species due to alignment issues and missing data.

Purpose of the Study:

• Introduce PolarBEAR, a novel method for inferring ancestral states using ancestral recombination graphs (ARGs).
• Evaluate PolarBEAR's accuracy in polarization and unfolded site-frequency spectrum (uSFS) estimation.
• Compare PolarBEAR with existing outgroup-based methods.

Main Methods:

• PolarBEAR infers ancestral states by utilizing local genealogies from reconstructed ARGs.
• Simulations were conducted under various scenarios to assess accuracy.
• The method was applied to human population data and compared against the est-sfs method.

Main Results:

• PolarBEAR demonstrates high accuracy in polarization and uSFS estimation, dependent on ARG reconstruction quality.
• It successfully polarizes sites missed by outgroup methods due to missing data.
• Results show high consistency with outgroup methods where applicable, with PolarBEAR estimating fewer high-frequency derived alleles.

Conclusions:

• PolarBEAR offers a robust alternative or complement to outgroup-based polarization methods.
• It is particularly valuable for species lacking suitable outgroup sequences.
• The method shows robustness across different mutation models and avoids biases seen in other methods with heterogeneous base composition.