Analysis of archaic human haplotypes suggests that 5hmC acts as an epigenetic guide for NCO recombination

Background Non-crossover (NCO) refers to a mechanism of homologous recombination in which short tracks of DNA are copied between homologue chromatids. The allelic changes are typically restricted to one or few SNPs, which potentially allow for the gradual adaptation and maturation of haplotypes. It is assumed to be a stochastic process but the analysis of archaic and modern human haplotypes revealed a striking variability in local NCO recombination rates. Methods NCO recombination rates of 1.9 million archaic SNPs shared with Denisovan hominids were defined by a linkage study and correlated with functional and genomic annotations as well as ChIP-Seq data from modern humans. Results We detected a strong correlation between NCO recombination rates and the function of the respective region: low NCO rates were evident in introns and quiescent intergenic regions but high rates in splice sites, exons, 5 '- and 3 '-UTRs, as well as CpG islands. Correlations with ChIP-Seq data from ENCODE and other public sources further identified epigenetic modifications that associated directly with these recombination events. A particularly strong association was observed for 5-hydroxymethylcytosine marks (5hmC), which were enriched in virtually all of the functional regions associated with elevated NCO rates, including CpG islands and 'poised' bivalent regions. Conclusion Our results suggest that 5hmC marks may guide the NCO machinery specifically towards functionally relevant regions and, as an intermediate of oxidative demethylation, may open a pathway for environmental influence by specifically targeting recently opened gene loci.

Automated summary

The NCO recombination rates show a strong correlation with the function of the region, with low rates in introns and quiescent intergenic regions and high rates in splice sites, exons, 5' and 3'-UTRs, as well as CpG islands. Epigenetic modifications such as 5hmC marks are associated with elevated NCO rates in functionally relevant regions, potentially allowing for environmental influence on gene loci.