Coding sequences are less susceptible to SV accumulation.

<p>(A) Genome browser view depicting the concordant and discordant coverage of the control (blue) and the tumor (red) samples across two protein-coding genes, and identified SVs therein. The detected insertions and deletions localize outside of coding sequences, and affect introns, intergenic spaces and UTRs. (B) Genome-wide breakpoint distribution across distinct functional compartments. Different subsets of the genome were selected according to following characteristics: <i>genome</i> corresponds to the full-length genome; the <i>unique genes</i> do not share common positions with any other gene; <i>overlapping genes</i> are non-unique genes; <i>exonic</i> regions, containing <i>3′UTRs, 5′UTRs</i> and coding sequences (<i>CDS</i>) were obtained from the unique genes in order to avoid ambiguity; In addition, <i>intronic</i> and <i>intergenic</i> regions as well as donor/acceptor splice sites (<i>splice sites</i>) were considered. For each subset the number of contained breakpoints was computed and normalized to the total length.