Correlation measures for linkage disequilibrium within and between populations

Area of Science:

• Population Genetics
• Statistical Genetics

Background:

• Linkage disequilibrium (LD) quantifies genetic association between loci.
• Measures like r2 (within populations) and rirj (between populations) are crucial for understanding population structure.

Purpose of the Study:

• To derive recurrence relationships for LD measures (r2 and rirj) under the island model of population subdivision.
• To validate these relationships using computer simulations and real-world data.

Main Methods:

• Developed recurrence relationships for r2 and rirj based on the linked identity-by-descent (LIBD) model.
• Utilized computer simulations to observe LD build-up after population subdivision.
• Applied derived equations to Hapmap data from African (YRI) and European (CEU) populations.

Main Results:

• The LIBD model accurately predicts the build-up and steady-state levels of r2 and rirj.
• Steady-state expectations for r2 depend on effective population size (Ne), number of subpopulations (k), and migration-recombination ratio (rho).
• Analysis of YRI and CEU populations suggests a separation time under 1000 generations, with ancient migration better explaining observed LD patterns.

Conclusions:

• The derived recurrence relationships provide a robust framework for studying LD in subdivided populations.
• Estimates of human population separation may need revision, considering the significant impact of ancient migration.
• The study highlights the utility of correlation statistics in inferring population history.