Linkage Disequilibrium

Linkage disequilibrium is the non-random association of alleles at different loci in a population: when particular alleles at nearby loci occur together more or less often than their individual frequencies would predict. It is a population-level pattern, distinct from family-level linkage, and it is the property that lets genome-wide association studies use a small set of markers to tag much larger stretches of the genome.

Definition

Linkage disequilibrium (LD) is the statistical association between alleles at two or more loci in a population, such that the frequency of a haplotype differs from the product of the constituent allele frequencies; it is quantified by measures such as D, D-prime, and r-squared.

Scope

The entry covers the definition and measures of linkage disequilibrium, the forces that create and erode it, its organization into haplotype blocks, and its use in association mapping of complex traits. It is a reference topic in population and medical genetics, not clinical guidance.

Core questions

• How does linkage disequilibrium differ from genetic linkage?
• What creates linkage disequilibrium and what breaks it down?
• How is it measured, and why does its structure form haplotype blocks?
• How do association studies exploit linkage disequilibrium to find disease loci?

Key concepts

• Non-random allele association
• Measures of LD (D, D-prime, r-squared)
• Haplotypes and haplotype blocks
• Recombination and the decay of LD
• Tag SNPs
• Genome-wide association study (GWAS)

Mechanisms

Linkage disequilibrium arises when a new allele appears on a particular chromosomal background and the two are inherited together, and it is reinforced by population history such as bottlenecks, admixture, drift, and selection. Recombination breaks down LD over generations: the closer two loci are, the fewer crossovers fall between them and the slower their association decays, so LD tends to be high for nearby loci and to diminish with distance. The pattern is summarized by measures including D, the normalized D-prime, and the correlation coefficient r-squared. Empirically the human genome is organized into haplotype blocks of high internal LD separated by recombination hotspots, as documented by Reich et al. (2001) and mapped systematically by the International HapMap Consortium (2005). Because alleles within a block are correlated, a few tag SNPs can capture most of the variation, which is the principle that genome-wide association studies use to scan the genome efficiently.

Clinical relevance

Linkage disequilibrium is what makes genome-wide association studies possible: an associated marker need not be causal but may simply be in LD with the true variant, so localizing a disease variant requires fine-mapping within the associated region. This entry describes how such evidence is generated and interpreted and is reference background, not a basis for individual diagnosis or treatment.

History

The idea of allelic association predates molecular markers, but its modern importance grew with dense human variation data. Reich et al. (2001) showed that LD in the human genome extends over substantial distances and is structured, work consolidated by the International HapMap Consortium (2005), which built a genome-wide map of common variation and its LD structure. Slatkin (2008) synthesized how LD records the evolutionary past and enables medical mapping, and reviews such as Hirschhorn and Daly (2005) set out how LD underpins genome-wide association studies of complex traits.

Debates

Does an LD-based association identify the causal variant?

Because association studies detect markers in LD with the true variant rather than the variant itself, a statistical association localizes a region but does not by itself prove which variant is causal; fine-mapping and functional study are needed to resolve this.

How does population structure confound association?

Allele-frequency differences between subpopulations can create association that mimics LD with disease, so methods that account for population structure are required to avoid spurious findings.

Key figures

• Montgomery Slatkin
• David Reich
• Jonathan Pritchard
• Joel Hirschhorn
• Mark Daly

Seminal works

• reich-2001
• hapmap-2005
• slatkin-2008

Frequently asked questions

How is linkage disequilibrium different from genetic linkage?

Linkage is the co-inheritance of loci within families across one or a few generations, whereas linkage disequilibrium is a population-level correlation of alleles built up and eroded over many generations; loci can be linked yet show little LD, and vice versa.

Why does linkage disequilibrium decay with distance?

Recombination over many generations breaks up associations between alleles, and because crossovers are more frequent between distant loci, LD is generally weaker the farther apart two loci are.