Fig. 4
From: Approximate search for known gene clusters in new genomes using PQ-trees
The outline of the algorithm that solves PQ-Tree Search. A During initialization the best derivations of the leaves of T are computed. The cell with two arrows marks the substitution between D and U. B PQ-Tree Alignment is solved for each substring of the target string S and the subtrees rooted in each internal node of T. C In this example, there are two derivations from the root of the PQ-tree; One is to the substring from index 1 to 6 and the other is to the entire string (if the character U is deleted from the string)