Application of Smodels in Quartet Based Phylogeny Construction

Abstract

Evolution is an important sub-area of study in biological science, where given a set of taxa, the goal is to reconstruct their evolutionary history, or phylogeny. One very recent approach is to predict a local phylogeny for every subset of 4 taxa, called a quartet topology, and then to assemble a phylogeny for the whole set of taxa satisfying these predicted quartet topologies. In general, the predicted quartet topologies might not always agree with each other, and thus the objective function becomes to satisfy a maximum number of them. This is the well known Maximum Quartet Consistency (MQC) problem. In the past, the MQC problem has been solved by dynamic programming and the so-called fixed-parameter method. Recently, we have proposed to solve the MQC in answer set programming. In this note, we summarize the theoretical results of this approach and report new experimental results for the purpose of comparison, which show that our approach in answer set programming is favored over the existing approaches based on dynamic programming and fixed-parameter method. In particular, some of the hard instances (where the error ratio is high) that were not reported to be solved in other approaches can now be solved in our approach.