Source-linked AI summary
SWEEPFINDER2: Increased sensitivity, robustness, and flexibility
Michael DeGiorgio, Christian D. Huber, Melissa J. Hubisz, Ines Hellmann, Rasmus Nielsen
TL;DR
Sweep detection based on polymorphism spectra can be confounded by background selection and other evolutionary forces. SweepFinder2 extends SweepFinder by modeling background selection and local recombination and incorporating invariant sites, providing improved power and flexibility for detecting selective sweeps.
Problem
Sweep footprints can resemble diversity reductions caused by background selection, motivating methods that account for this confounding force when identifying positive selection.
Method
SweepFinder2 modifies neutral frequency spectra using local B-values, includes invariant sites, and compares a scaled-spectrum null model with a sweep alternative that accounts for local recombination.
Results
The generalized composite likelihood ratio test provides a substantial improvement in power and flexibility over SweepFinder while accounting for negative selection when searching for adaptive alleles.
Takeaways & Limitations
SweepFinder2 supports sweep detection that distinguishes selective-sweep signals from diversity reductions associated with background selection.
Abstract
from arXiv · showhide
SweepFinder is a popular program that implements a powerful likelihood-based method for detecting recent positive selection, or selective sweeps. Here, we present SweepFinder2, an extension of SweepFinder with increased sensitivity and robustness to the confounding effects of mutation rate variation and background selection, as well as increased flexibility that enables the user to examine genomic regions in greater detail and to specify a fixed distance between test sites. Moreover, SweepFinder2 enables the use of invariant sites for sweep detection, increasing both its power and precision relative to SweepFinder.
Introduction
Positive-selection footprints can be confounded by background selection and other evolutionary forces, motivating sweep-detection methods that account for these effects. This matters because background selection is widespread and can reduce diversity in ways that resemble selective sweeps.
- Background selection is the loss of neutral variation caused by purging linked deleterious alleles through negative selection.
- Background selection is widespread in humans, with estimated mean diversity reductions of 19–26% on autosomes and 12–40% on the X chromosome.
- Selective sweeps can reduce local diversity and diversity relative to divergence across megabase-scale genomic regions.
- Because background selection can mimic sweep-associated diversity reductions, diversity-only methods may falsely infer positive selection.
- These confounding signals motivate methods that identify sweeps while accounting for background selection, especially in regions of low recombination.
SweepFinder2
SweepFinder2 extends SweepFinder by modeling background selection and local recombination while incorporating invariant sites into sweep detection. The resulting composite likelihood ratio test improves power and flexibility for detecting selective sweeps.
- SweepFinder2: SweepFinder2 modifies the neutral derived frequency spectrum using local B-values and includes invariant sites, specifically substitutions.
- SweepFinder2: B-values range from 0 to 1 and represent local reductions in genetic diversity or effective population size due to background selection.
- SweepFinder2: The test uses the locally scaled spectrum under the null hypothesis and a sweep model accounting for local recombination rate under the alternative hypothesis.
- SweepFinder2: Regions with low B-values provide little sweep evidence when null and alternative frequency spectra are similar, whereas relatively high B-values can retain sweep signals.
- SweepFinder2: SweepFinder2 accounts for negative selection when searching for adaptive alleles and generalizes SweepFinder’s composite likelihood ratio test.
- SweepFinder2: The generalized test provides a substantial improvement in power and flexibility over the widely used SweepFinder software.