Fig. 4.

Investigation of tandem repeats to assess merging strategies’ performance. a Illustration of a locus where eight insertion alleles (Input) have between + 2 and + 5 copies of a 29-bp repeat across 10 samples. Four of the alleles are annotated as redundant representations (blue) since they have a counterpart with an equal number of copies (orange). A correct merge would preserve each of the unique alleles and remove all redundant alleles, leaving 0 missing and 0 redundant SVs in the locus. An incorrect merge removes two unique insertions (+ 3, + 4) and leaves 1 redundant insertion. b Boxplot of the number of missing variants per locus for each merging strategy. c Barplot of the number of loci with none or any redundant alleles post-merging