Sequence features in regions of weak and strong linkage disequilibrium

Abstract

We use genotype data generated by the International HapMap Project to dissect the relationship between sequence features and the degree of linkage disequilibrium in the genome. We show that variation in linkage disequilibrium is broadly similar across populations and examine sequence landscape in regions of strong and weak disequilibrium. Linkage disequilibrium is generally low within approximately 15 Mb of the telomeres of each chromosome and noticeably elevated in large, duplicated regions of the genome as well as within approximately 5 Mb of centromeres and other heterochromatic regions. At a broad scale (100-1000 kb resolution), our results show that regions of strong linkage disequilibrium are typically GC poor and have reduced polymorphism. In addition, these regions are enriched for LINE repeats, but have fewer SINE, DNA, and simple repeats than the rest of the genome. At a fine scale, we examine the sequence composition of "hotspots" for the rapid breakdown of linkage disequilibrium and show that they are enriched in SINEs, in simple repeats, and in sequences that are conserved between species. Regions of high and low linkage disequilibrium (the top and bottom quartiles of the genome) have a higher density of genes and coding bases than the rest of the genome. Closer examination of the data shows that whereas some types of genes (including genes involved in immune response and sensory perception) are typically located in regions of low linkage disequilibrium, other genes (including those involved in DNA and RNA metabolism, response to DNA damage, and the cell cycle) are preferentially located in regions of strong linkage disequilibrium. Our results provide a detailed analysis of the relationship between sequence features and linkage disequilibrium and suggest an evolutionary justification for the heterogeneity in linkage disequilibrium in the genome.

Figure 1.

Pairwise disequilibrium coefficients (r²) for one window in the genome. Tan circles denote the observed values. Green line denotes the average of observed values. Red line denotes the curve resulting from the fitted model, which models the decay of linkage disequilibrium as a function of the per base-pair population recombination rate, 4Nρ, and the distance between markers d_ij (see Methods). The example refers to marker pairs in the window from 2–3 Mb on chromosome 3.

Figure 2.

Genome-wide summary of fitted linkage disequilibrium values and identified “hotspots” for the rapid breakdown of linkage disequilibrium. (Top) The fitted disequilibrium coefficients for markers separated by 30 kb. Disequilibrium coefficients were calculated within 100-kb windows distributed throughout the genome. (Bottom) Intermarker intervals (in red), where linkage disequilibrium decays very rapidly, such that disequilibrium between spanning marker pairs is generally low (for details, see text). Evaluated intervals where disequilibrium did not appear to decay very rapidly are marked in light blue.

Figure 3.

Genome-wide summary of fitted linkage disequilibrium values and identified “hotspots” for the rapid breakdown of linkage disequilibrium. (Top) The fitted disequilibrium coefficients for markers separated by 30 kb. Disequilibrium coefficients were calculated within 100-kb windows distributed throughout the genome. (Bottom) Intermarker intervals (in red), where linkage disequilibrium decays very rapidly, such that disequilibrium between spanning marker pairs is generally low (for details, see text). Evaluated intervals where disequilibrium did not appear to decay very rapidly are marked in light blue.

Figure 4.

Genome-wide summary of fitted linkage disequilibrium values and identified “hotspots” for the rapid breakdown of linkage disequilibrium. (Top) The fitted disequilibrium coefficients for markers separated by 10 kb. Disequilibrium coefficients were calculated within 100-kb windows distributed throughout the genome. (Bottom) Intermarker intervals (in red), where linkage disequilibrium decays very rapidly, such that disequilibrium between spanning marker pairs is generally low (for details, see text). Evaluated intervals where disequilibrium did not appear to decay very rapidly are marked in light blue.

Figure 5.

Variation of fitted linkage disequilibrium values (for markers separated by 30,000 bp) across the three groups of samples and of selected sequence features including sequence polymorphism, total repeat content, GC content, proportion of coding bases, and proximity to the centromeres. Results refer to 1000-kb windows across chromosome 3.