Figure 2
From: CLUSS: Clustering of protein sequences based on a new similarity measure
Merging leaves. Let us take a rooted phylogenetic tree with L 1 , L 2 ...L 10 as leaves, and N 1 , N 2 ...N 9 as internal nodes, where N 5 and N 9 are identified as low co-similarity nodes (black nodes). Leaves are merged until a black node is reached, except for L 3 , L 4 , L 5 and L 6 , which need special consideration. All leaves connected between N 5 and N 9 are merged into a distinct subfamily. L 3 is connected directly to N 5 so it constitutes a distinct subfamily. We thus obtain the subfamilies Sf1, Sf2, Sf3 and Sf4, while Sf2 contains the orphan sequence represented by leaf L3.