A novel class of non-coding variants driving DNA double-strand breaks is associated with complex genetic diseases

Previous research efforts to map the genetic determinants of complex diseases genome-wide have identified genetic variants that act at the protein level by disrupting function, or at the regulatory level by affecting gene expression. However, for most mutations, the underlying molecular mechanisms remain unknown. In addition to protein function and gene expression level, complex diseases have previously been associated with DNA double-strand breaks (DSBs), particularly in the case of cancer, immune disorders, and neurological and psychiatric diseases. Such DSBs have the potential to regulate or interfere with transcription, replication, and genome maintenance. However, no mutations that alter the occurrence of DSBs have yet been reported. By mapping genome-wide SNPs that alter the binding of DNA repair proteins, we discovered a new class of non-coding SNPs that we named dsbSNPs and which modulate DSB frequency. Importantly, while some DSBs occur in transcribed chromatin, transcription does not seem to have a causal effect on DSBs. Instead, dsbSNPs can influence the selection of the repair pathway depending on whether they alter transcription factor binding or not, thereby potentially controlling the mutation rate in the vicinity of DSBs. dsbSNPs showed high enrichment for GWAS top associations, twice as high as for eSNPs. Such associations support a novel mechanism by which non-coding variants altering DSB formation may play a role in complex genetic diseases.