Measurement of the similarity between classification trees. (a) Computation of the traveling distance between leaves. In a classification tree, between any two species there exists a lowest common ancestor (LCA) and, a species can be considered as a leaf. Given a pair of leaves X and Y, if a worm starts from X and travels along the shortest path between the two leaves while another worm starts from Y and travels along the same path, before they meet at the LCA node, the first worm needs to go through m ancestor nodes of X (AXm) and the other worm needs to go through n ancestor nodes of Y (AYm). In this example, the traveling distance between X and Y is calculated as m + n. (b) Similarity between some sample trees. A tree can be transformed into a distance matrix after computing all the leaf-to-leaf traveling distances. This figure exhibits the distance matrices of four sample trees. The topological similarity between two trees is then computed as the correlation coefficient between the distance matrices of the trees; a high correlation coefficient indicates a high similarity.
Lee et al. BMC Genomics 2012 13:157 doi:10.1186/1471-2164-13-157