Quartet-based phylogenetic inference: A grid approach

Abstract

The accuracy of quartet-puzzling method, which is widely used in molecular phylogenetic analysis depends heavily on the number of intermediate trees searched and the order of molecular sequences selected to generate these trees. Because the quality of intermediate trees cannot be guaranteed, the consensus results can easily be trapped into local optima. In this paper we present a new approach to guide the intermediate tree selection. Our experimental results show that the accuracy of reconstructed trees can be improved significantly. Using our method, the task can easily be partitioned into independent subtasks of different sizes. Therefore, it can effectively be implemented and run in the heterogeneous and dynamic environment of the computational grid. © Springer-Verlag Berlin Heidelberg 2004.