Fig. 2.

Distribution of the separation between minimizers for k = 7 and w = 11 on DNA sequences. (A) Results on a de Bruijn sequence of order w + k. (B) Results computed on the human reference genome (hg19). Each line represents a different minimizer scheme using a different ordering. Note that previous heuristic orderings all behave like the randomized orderings (uniform distribution) except for separation of 1 and 2. The universal k-mer ordering computed by DOCKS has a noticeably different distribution, with a mode and a higher mean