Characterization of 26 deletion CNVs reveals the frequent occurrence of micro-mutations within the breakpoint-flanking regions and frequent repair of double-strand breaks by templated insertions derived from remote genomic regions.

Copy number variations (CNVs) have increasingly been reported to cause, or predispose to, human disease. However, a large fraction of these CNVs have not been accurately characterized at the single-base-pair level, thereby hampering a better understanding of the mutational mechanisms underlying CNV formation. Here, employing a composite pipeline method derived from various inference-based programs, we have characterized 26 deletion CNVs [including three novel pathogenic CNVs involving an autosomal gene (EXT2) causing hereditary osteochondromas and an X-linked gene (CLCN5) causing Dent disease, as well as 23 CNVs previously identified by inference from a cohort of Canadian autism spectrum disorder families] to the single-base-pair level of accuracy from whole-genome sequencing data. We found that breakpoint-flanking micro-mutations (within 22 bp of the breakpoint) are present in a significant fraction (5/26; 19%) of the deletion CNVs. This analysis also provided evidence that a recently described error-prone form of DNA repair (i.e., repair of DNA double-strand breaks by templated nucleotide sequence insertions derived from distant regions of the genome) not only causes human genetic disease but also impacts on human genome evolution. Our findings illustrate the importance of precise CNV breakpoint delineation for understanding the underlying mutational mechanisms and have implications for primer design in relation to the detection of deletion CNVs in clinical diagnosis.