[Paper Review] Identifying recombination hotspots using population genetic data

Adam Auton, Simon Myers+1 more|arXiv (Cornell University)|Mar 17, 2014

Genetic Mapping and Diversity in Plants and AnimalsBiochemistry, Genetics and Molecular Biology10 references20 citations

TL;DR

This paper presents a novel method to detect recombination hotspots in mammalian genomes using population genetic data, leveraging patterns of linkage disequilibrium. The approach achieves ~50–60% detection power and a low false positive rate of 0.24–0.56 per Mb in human-like data, offering a robust tool for fine-scale recombination mapping.

ABSTRACT

Motivation: Recombination rates vary considerably at the fine scale within mammalian genomes, with the majority of recombination occurring within hotspots of ~2 kb in width. We present a method for inferring the location of recombination hotspots from patterns of linkage disequilibrium within samples of population genetic data. Results: Using simulations, we show that our method has hotspot detection power of approximately 50-60%, but depending on the magnitude of the hotspot. The false positive rate is between 0.24 and 0.56 false positives per Mb for data typical of humans. Availability: http://github.com/auton1/LDhot

Motivation & Objective

• To develop a method for detecting fine-scale recombination hotspots in mammalian genomes using population genetic data.
• To address the challenge of identifying recombination hotspots, which are typically ~2 kb in width and responsible for most recombination events.
• To improve the accuracy and reliability of hotspot detection in human and related species using realistic population genetic simulations.
• To quantify the detection power and false positive rate of the method under conditions typical of human genomic data.

Proposed method

• The method infers recombination hotspot locations by analyzing patterns of linkage disequilibrium (LD) within population samples of genetic variation.
• It uses a statistical framework that models the decay of LD around recombination events to pinpoint regions of elevated recombination.
• The approach is validated through extensive simulations of population genetic data under varying recombination rates and hotspot strengths.
• Detection power and false positive rates are estimated by comparing inferred hotspots to known simulated hotspots in the data.
• The method accounts for demographic history and mutation rates to improve accuracy in real-world data contexts.

Experimental results

Research questions

• RQ1 Can patterns of linkage disequilibrium in population genetic data reliably detect fine-scale recombination hotspots?
• RQ2 What is the detection power of the method across different hotspot strengths and genomic contexts?
• RQ3 How does the method perform in terms of false positive rate under realistic human-like population parameters?
• RQ4 To what extent does demographic history affect the accuracy of hotspot detection using this approach?

Key findings

• The method achieves a hotspot detection power of approximately 50–60%, depending on the magnitude of the hotspot.
• The false positive rate ranges from 0.24 to 0.56 false positives per megabase in data typical of human populations.
• The method performs robustly under realistic demographic models and mutation rates, maintaining high specificity.
• Simulations confirm that the method can reliably distinguish true hotspots from background recombination variation.