Extended Data Fig. 4: Duplications within direct repeats.

a, Examples of duplications within direct repeats. Reference sequence shown. Blue text indicates location of repeat motifs. Blue highlight shows location of motif mismatches. Orange highlight indicates region duplicated in gnomAD. Orange text with blue highlight indicates that the duplication includes the interruption. b, Duplication position bias. Duplication start/end positions (blue) or positions flanking duplications (orange), categorized by their location (ie. start of left motif, mismatch position in left motif, end of spacer, etc.). Frequency of the duplication at each position (observed) is divided by the portion of the motif for which the position accounts (expected). Represents n = 3170 DR loci with spacer length < 10nt and containing a duplication >5 nt. c) Gap repair model explaining direct repeat duplications. The initial A-B-A pattern forms a slipped-strand structure. Post-replicative gap-filling (Polζ) fills in single-stranded loop regions, resulting in 4-way pseudo-Holliday junctions. Cleavage of the top strand allows the filled-in sequence to be ligated into the top strand. Either this can be repeated for the other loop, or replication of the top strand will produce a daughter cell with the A-B-A-B-A pattern.