Ancestry inference in complex admixtures via variable-length Markov chain linkage models.

Area of Science:

• Population Genetics
• Genomics
• Bioinformatics

Background:

• Accurate inference of ancestral chromosomal segments is crucial for genetic studies, including population demographics, history, and disease gene mapping.
• Existing methods for ancestry inference have limitations, relying on either simplified linkage disequilibrium (LD) models or complex models requiring explicit ancestral haplotypes.

Purpose of the Study:

• To introduce ALLOY, an efficient and accurate method for inferring the ancestral origins of chromosomal segments in admixed individuals.
• To improve upon existing ancestry inference techniques by incorporating generalized and expressive linkage disequilibrium models.

Main Methods:

• ALLOY utilizes a factorial hidden Markov model to represent the distinct maternal and paternal admixture processes.
• It models background linkage disequilibrium in ancestral populations using an inhomogeneous variable-length Markov chain.
• The method was evaluated across diverse admixture scenarios, including recent and ancient admixtures with up to four ancestral populations.

Main Results:

• ALLOY demonstrates superior performance compared to previous state-of-the-art methods in ancestry inference.
• The method shows robustness even when faced with uncertainties in model parameters.
• It effectively handles a range of admixture histories and population numbers.

Conclusions:

• ALLOY provides a powerful and efficient tool for inferring ancestral origins in admixed populations.
• Its advanced modeling of linkage disequilibrium enhances accuracy in genetic analyses.
• The method's robustness makes it reliable for diverse population genetic research applications.