Abstract

Phylogenetic (evolutionary) tree estimation is fundamental to biology, and has applications to much biomedical research. Phylogenetic tree estimation is enormously difficult: the best approaches are NP-hard, and many real datasets take weeks or months of analysis. In particular, the estimation of multiple sequence alignment turns out to be one of the most challenging and important steps in a phylogenetic analysis.

In this talk I will present several methods that produce greatly improved trees as compared to other phylogeny estimation methods. These methods utilize divide-and-conquer strategies in order to improve the accuracy of traditional phylogeny estimation methods. Among the methods I will present are SATe (Liu et al. 2009, Science Vol 324, no. 5934) and DACTAL (in preparation). SATe simultaneously estimates a tree and a multiple sequence alignment, while DACTAL estimates a tree without ever computing an alignment on the entire set of sequences. Both methods provide great improvements in tree accuracy over other methods, and do so fairly efficiently.