The lengths of admixture tracts

Area of Science:

• Population Genetics
• Genomic History
• Computational Biology

Background:

• Admixture tract length distribution is key for inferring past gene flow timing.
• Current methods often assume independently and identically distributed (iid) exponential tract lengths.
• This assumption is foundational for hidden Markov model (HMM) based admixture analyses.

Purpose of the Study:

• To evaluate the validity of the iid exponential tract length assumption.
• To compare tract length distributions under various population-genetic models.
• To develop improved methods for analyzing admixture tract lengths, particularly for recent admixture.

Main Methods:

• Comparison of expected admixture tract length distributions from different population-genetic models.
• Analysis under the Wright-Fisher model with recombination.
• Development of a dyadic interval-based stochastic process for generating admixture tracts.

Main Results:

• The iid exponential tract length assumption is violated for both recent and ancient admixture events under the Wright-Fisher model.
• Reliance on this assumption can result in false positives when estimating the number of admixture events.
• Existing models show poor performance for populations with recent admixture.

Conclusions:

• The standard assumption of iid exponential admixture tract lengths is not universally applicable.
• Inaccurate assumptions can lead to misinterpretations of population history and gene flow.
• A novel dyadic interval-based stochastic process offers a more accurate approach for analyzing recent admixture scenarios.