A Likelihood Approach for Uncovering Selective Sweep Signatures from Haplotype Data

Area of Science:

• Population Genetics
• Evolutionary Biology
• Genomics

Background:

• Selective sweeps are key genomic signatures of natural selection and adaptation.
• Existing methods for detecting sweeps from haplotype data lack a model-based approach.
• Identifying sweeps is crucial for understanding evolutionary adaptation mechanisms.

Purpose of the Study:

• To develop a novel model-based statistical framework for detecting selective sweeps.
• To introduce a likelihood ratio test statistic (T) for identifying sweep signatures in whole-genome data.
• To infer the number of concurrently sweeping haplotypes within a population.

Main Methods:

• Developed a likelihood ratio test statistic (T) based on a haplotype frequency spectrum distortion model.
• Applied the T statistic to whole-genome polymorphism data from human populations (CEU, YRI) and Drosophila melanogaster.
• Validated the method's ability to detect hard and soft sweeps across diverse demographic scenarios and selection parameters.

Main Results:

• The T statistic successfully identified known and novel selective sweep candidates in human populations (e.g., LCT, HLA genes) and Drosophila (e.g., Ace).
• The method demonstrated robustness in detecting both hard and soft sweeps under various conditions.
• Open-source software for computing the T statistic and haplotype counts is provided.

Conclusions:

• The proposed T statistic offers a powerful, model-based approach for detecting selective sweeps.
• This method enhances the ability to study genetic adaptation across diverse species.
• The tool facilitates further research into the genetic basis of adaptation in natural populations.