We have shown that rates of mutation (intergenic, intronic and ancient repeat divergence as well as SNP density) and synonymous selection are correlated with chromatin structure. Regions of open chromatin display the lowest mutation rates and the least constraint at the synonymous sites of genes. Consequently previous observations of mutational hotspots in the human genome, high mutation rates around classes of genes involved in extracellular communication, the low dN/dS observed in housekeeping genes and the clustering of genes with similar divergence levels can all also be associated with chromatin structure. These correlations are observed despite the relatively low resolution of the chromatin dataset. The average length of the clones used in this analysis was 146 kb but the average human exon is approximately a thousand times smaller than this. There is consequently a disparity between the DNA regions whose rate of change we are measuring and the regions whose chromatin structure is known. The ability to measure chromatin structure at a higher resolution in the future may help increase the strength of these observed correlations.
We believe the lower background mutation rate observed in open regions of the genome in this study is likely to be a result of these regions being more accessible to repair mechanisms. Indeed it is known that sites of transcription-coupled repair are clustered in the gene dense (and therefore) open chromatin regions of the genome [21], that chromatin remodelling is a precursor to DNA repair, and that efficient DNA lesion detection is associated with relaxed chromatin structures [22-24]. However, contrary to mutation rate, we believe it unlikely that chromatin structure mediates selection on synonymous sites directly. Rather, it is more likely that genes that display a high level of selection at their synonymous sites are preferentially located in closed regions of the genome. It may be that these genes in general require especially tight transcriptional regulation, with a consequence being they are less accessible for DNA repair.
Chromatin structure is likely, however, to be only one of a number of factors that are associated with the variance in divergence rates observed across the human genome. This is supported by the fact that the levels of intergenic divergence of chromosome 19 are substantially higher than other autosomes, despite being gene dense and relatively open in structure. Most notably, both the chromatin dataset used in this analysis, as well as nucleosome formation potential [10], have previously been associated with GC content. Although this agreement between the lymphoblastoid chromatin dataset used in this analysis and other more general datasets is reassuring, GC content has previously been associated with rates of mutation and selection. However, although the mechanisms underlying the appearance of GC variability and isochores along the human genome remain controversial, it has been proposed that they may be a result of selection for the structural requirements of DNA. For example, an increase in GC content has been associated with an increase in bendability of DNA and a decrease in curvature, properties associated with more open chromatin [25]. Further analysis is consequently required to determine the complex interplay between the various factors involved in rates of mutation and selection across the human genome.
Answers to all your Biology Questions
