Figure 3

Two phylogenetic trees are shown, which were built from the same protein sequence alignment, but that correspond to two different domains from the Pfam database [21]: TBC and UCH. This example has already been described in the literature [24] and is used here as a benchmark for PhyloPattern. The results of each step in PhyloPattern (based on the strategy described in Pattern Matching) are shown. Step 1, Annotation: the full domain architecture is given for each sequence. Domain architectures for internal nodes are computed with the Dollo parsimony algorithm [18]. Step 2, Pattern Matching: the pattern shown above each tree is used to detect "parent" nodes of a shuffling event resulting in the architecture TBC-UCH (indicated in yellow). Step 3, Annotation: a purple circle is placed on the derived branch to locate the event. Step 4, Pattern Matching: A simple pattern is applied to extract leaves from each derived subtree. The human sequence ENSP305473 is common to the two subtrees and can be used as a reference for a subsequent genomic comparison with sequences having the "parent" architecture. Labeled trees for TBC and UCH domains are provided [see Additional file 4], [see Additional file 5].