An experimental study of quartets MaxCut and other supertree methods

Abstract

Although many supertree methods have been developed in the last few decades, none has been shown to produce more accurate trees than the popular Matrix Representation with Parsimony (MRP) method. In this paper, we evaluate the performance of several supertree methods based upon the Quartets MaxCut method of Snir and Rao. We show that two of these methods usually outperform MRP and all other supertree methods we studied under many realistic model conditions. In addition, we show that the popular criterion of minimizing the total topological distance to the source trees is only weakly correlated with topological accuracy, and therefore that evaluating supertree methods on biological datasets is problematic. © 2010 Springer-Verlag.