A Genetic Algorithm Formulation for Rogue Taxa Problem

Phylogenetic reconstruction methods like the most common maximum likelihood or Bayesian methods usually result in a large number of slightly varying trees over same set of taxa, known as bootstrap replicates (usually hundreds to thousands in numbers). Extracting useful information from this set of trees involve summarizing most common information in all these trees in a single tree, often called a consensus tree. Presence of rogue taxa greatly reduces resolution of consensus. Some good algorithms are available to identify and remove these but they work only for binary trees. This paper flows with the rogue taxa problem, description and comparison of available algorithms, and our Genetic Algorithm formulation for the problem.