Efficient and Robust Analysis of Large Phylogenetic Datasets

Abstract

The goal of phylogenetics is to reconstruct ancestral relationships between different taxa, e.g., different species in the tree of life, by means of certain characters, such as genomic sequences. We consider the prominent problem of reconstructing the basal phylogenetic tree topology when several subclades have already been identified or are well known by other means, such as morphological characteristics. Whereas most available tools attempt to estimate a fully resolved tree from scratch, the profile neighbor-joining (PNJ) method focuses directly on the mentioned problem and has proven a robust and efficient method for large-scale datasets, especially when used in an iterative way. We describe an implementation of this idea, the ProfDist software package, which is freely available, and apply the method to estimate the phylogeny of the eukaryotes. Overall, the PNJ approach provides a novel effective way to mine large sequence datasets for relevant phylogenetic information.