Human mutation rate associated with DNA replication timing

Abstract

Eukaryotic DNA replication is highly stratified, with different genomic regions shown to replicate at characteristic times during S phase. Here we observe that mutation rate, as reflected in recent evolutionary divergence and human nucleotide diversity, is markedly increased in later-replicating regions of the human genome. All classes of substitutions are affected, suggesting a generalized mechanism involving replication time-dependent DNA damage. This correlation between mutation rate and regionally stratified replication timing may have substantial evolutionary implications.

FIGURE 1. Replication time-dependence of evolutionary divergence and human SNP density

Shown are replication time-dependence of (a–c) human-chimpanzee divergence (fraction sites changed); (d–f) human-macaque divergence; and (g–i) human polymorphism rate (SNP density), computed across 44 regions (ENCODE) comprising 1% of the human genome. Analyses are presented for all putatively neutral sites (a,d,f – left column), for neutral sites with all CpG-prone dinucleotides removed (b,e,h – middle column), and for CpG dinucleotides only (c,f,i – right column). CpG-prone dinucleotides were defined as all sites immediately preceded by C of followed by G in either species, CpG dinucleotides were defined as sites for which C was immediately followed by G at least in one of the species. Plots show mutation rates averaged over all 50kb non-overlapping windows within each of four temporal replication states (S1–S4), together with 95% confidence intervals for the mean. Significance of each trend is shown within the corresponding panel as a p-value (Cochran-Armitage trend test for proportions). For divergence analysis, human-chimpanzee hg17vsPanTro1 and human-macaque hg17vsRheMac2 axtNet alignments were processed to exclude all regions with more than a single substitution per sliding 5-nucleotide window and axtNet fragments which were either shorter than 500 bases or demonstrated average substitution rate above 3% (12% in case of macaque) were also discarded. For polymorphism analysis, we used Version 1 bulk SNPs data set published as part of the Personal Genome Sequence project (Watson).