Figure 5.

Sequence preferences around human deletions. (a) Nucleotide profiles. Shown are profiles of log odds for the nucleotide composition around human deletions of lengths 1-4 (Figure 3; see Materials and methods). The 0.95 confidence intervals are plotted for each frequency. For each offset relative to the deletion point, we computed the frequency of each nucleotide and compared it to the frequency in genomic regions with similar GC content. Shaded regions represent the nucleotide preferences of the deleted sequence itself. The analogous chimp profiles are shown in Figure S4 in Additional data file 1. (b) Higher order nucleotide correlation. Shown are schematic illustrations of the two strongest statistical associations between pairs of nucleotides flanking insertions or deletions of a specific length. Both cases represent co-occurrence of G-C pairs flanking a deletion of even size (red nucleotides represent the deleted fragment). It is possible that base pairing between these two positions contributes to specific deletion scenarios.