Comparing arbitrary unrooted phylogenetic trees using generalized matching split distance

In the paper, we describe a method for comparing arbitrary, not necessary fully resolved, unrooted phylogenetic trees. The proposed method is based on finding a minimum weight matching in bipartite graphs and can be regarded as a generalization of well-known Robinson-Foulds distance. We present some properties and advantages of the new distance. We also investigate some properties of the presented distance in a common biological problem of finding a single phylogenetic tree (consensus tree) that reliably represents a set of various phylogenetic trees. Finding a consensus tree (or a small set of such trees) is an important phase in the phylogenetic research, especially if a method that is chosen for the construction process returns a set of trees (for example the very popular Bayesian approach).