Population genetics · BEGINNER LESSON
Out-of-sample LD
In these sources, linkage disequilibrium (LD) is the nonrandom association of alleles at different loci, meaning the alleles are found together more or less often than expected from their separate frequencies. Because the phrase “out-of-sample LD” is not defined in the excerpts, this lesson uses it to focus on LD as it is observed and estimated from a sample, where sampling and population structure can change the measured value.
What LD is and what is measured
LD is a description of how alleles at different loci are associated. If two alleles, A at one locus and B at another, are independent, the AB haplotype would be expected at the product of their separate frequencies. A common measure is D, defined as pAB − pA pB, the gap between the observed joint frequency and the random-combination expectation. Another common summary is r^2, which rescales this departure and is often used because it is easy to compare across data sets, although it is only defined when both loci are segregating.
[1] One Hundred Years of Linkage Disequilibrium - PMC [2] Linkage disequilibrium — understanding the evolutionary past and ... [3] Linkage Disequilibrium in Humans: Models and Data - PMC
How LD arises and changes
Several processes can create or alter LD. Recombination, the exchange of chromosome segments during meiosis, tends to reduce D from one generation to the next by a factor related to the recombination rate c. Genetic drift, which is random change in allele frequencies from one generation to the next, can create LD in small samples or small populations. Mixing subpopulations with different allele frequencies also creates LD in the combined sample, even when each subpopulation separately shows no LD. Because of this, LD is often strongest for closely linked loci and can also appear for unlinked loci when population structure is present.
[1] One Hundred Years of Linkage Disequilibrium - PMC [2] Linkage disequilibrium — understanding the evolutionary past and ... [3] Linkage Disequilibrium in Humans: Models and Data - PMC
Why sample-based LD can be misleading
The measured value of r^2 for unlinked loci can be strongly affected by sample size, so small samples may show apparent LD even when the true long-term association is weak. Population substructure, inbreeding, bottlenecks, and admixture can inflate LD and produce false positives in association studies. For that reason, LD statistics are descriptive: they quantify nonrandom association, but by themselves they do not explain the cause or prove physical linkage.
[1] One Hundred Years of Linkage Disequilibrium - PMC [2] Linkage disequilibrium — understanding the evolutionary past and ... [3] Linkage Disequilibrium in Humans: Models and Data - PMC
Limits of the measures
No single LD statistic is best in every situation. D depends on allele frequencies, so its possible range changes from one locus pair to another. r^2 is popular but can be hard to use for exact multi-generation calculations and can be dominated by sample size for unlinked loci. Therefore, interpreting out-of-sample LD requires attention to the marker set, the number of sampled individuals, and whether the sample mixes populations.
[1] One Hundred Years of Linkage Disequilibrium - PMC [2] Linkage disequilibrium — understanding the evolutionary past and ... [3] Linkage Disequilibrium in Humans: Models and Data - PMC
SEE THE IDEA
How a sampled LD value is formed
Start with how often each allele appears at each locus.
Count how often the two alleles appear together on the same chromosome copy.
Compare observed joint frequency with random expectation using D or r^2.
Recombination, drift, and mixing can raise or lower the measured value.
The measured LD value is the result of both biological history and how the sample was collected.
A sample can show nonrandom allele association for biological reasons, sampling reasons, or both.
TRY IT YOURSELF
Test your understanding
Choose an answer, then check the explanation.
PUT IT TOGETHER
Match each concept to its meaning
Drag a label onto its description, or select a label and then select a description. This also works with a keyboard.
4 pairs to match
Sources and further reading
This general lesson was generated from public source material. It explains a concept; it does not interpret a private paper or provide medical advice.