Large-scale correlations between eight key double strand break related data sets over the whole human genome

Abstract

Eight different data sets, covering the whole human genome are compared with regard to their genomic distribution. A close correlation between cytological detected chiasma and MLH1 immunofluorescence sites with the recombination density distribution from the HapMap project was found. Sites with a high probability of chromatid breakage after exposure to low and high ionization density radiations are often located inside common and rare Fragile Sites (FSs) indicating that the common Radiation-Induced Breakpoint sites (RIBs) may be a new kind of more local fragility. Furthermore, Oncogenes and other cancer-related genes are commonly located in regions with an increased probability of rearrangements during genomic recombination, or in regions with high probability of copy number changes, possibly since these processes may be involved in oncogene activation and cancer induction. An increased CpG density is linked to regions of high gene density to secure high fidelity reproduction and survival. To minimize cancer induction these genes are often located in regions of decreased recombination density and/or higher than average CpG density. Interestingly, copy number changes occur predominantly at common RIBs and/or FSs at least for breast cancers with poor prognosis and they decrease weakly but significantly in regions with increasing recombination density and CpG density. It is compelling that all these datasets are influenced by the cells handling of double strand breaks and more generally DNA damage on its genome. In fact, the DNA repair genes are systematically avoiding regions with a high recombination density. This may be a consequence of natural selection, as they need to be intact to accurately handle repairable DNA lesions.