Haplotype-Based Models Improve Sweep Detection in Ancient Populations with Complex Demography.

Area of Science:

• Population Genetics
• Ancient DNA Analysis
• Evolutionary Biology

Background:

• Identifying positive selection in human genomes is complicated by demographic factors like admixture and migration.
• Traditional methods often rely on allele frequency or site frequency spectrum (SFS) data, with limited use of haplotype information, especially for ancient DNA.

Purpose of the Study:

• To evaluate the performance of haplotype-based and SFS-based methods for detecting selective sweeps in ancient European populations.
• To adapt and assess the saltiLASSI haplotype framework for pseudohaploid ancient genomes.

Main Methods:

• Extended the saltiLASSI haplotype-based likelihood framework to handle pseudohaploid ancient genomes, utilizing truncated haplotype frequency spectra and spatial decay.
• Conducted forward-in-time simulations incorporating varying sweep ages, admixture events, and selection dynamics.
• Compared the performance of saltiLASSI against the SFS-based method SweepFinder2.

Main Results:

• Haplotype-based likelihood models, particularly saltiLASSI, demonstrated superior power in detecting selective sweeps within admixed populations compared to SFS methods.
• Haplotype methods were more effective when sweep haplotypes originated from migration or when selection's signature was recent post-admixture.
• The adapted saltiLASSI framework successfully utilized truncated haplotype data from ancient genomes.

Conclusions:

• Haplotype-based inference holds significant promise for analyzing ancient DNA and improving the detection of historical selective sweeps.
• Model-based approaches, like saltiLASSI, are crucial for accurately identifying selection signatures in populations with complex demographic histories.