Online Consensus and Agreement of Phylogenetic Trees

Computational heuristics are the primary methods for reconstruction of phylogenetic trees on large datasets. Most large-scale phylogenetic analyses produce numerous trees that are equivalent for some optimization criteria. Even using the best heuristics, it takes significant amount of time to obtain optimal trees in simulation experiments. When biological data are used, the score of the optimal tree is not known. As a result, the heuristics are either run for a fixed (long) period of time, or until some measure of a lack of improvement is achieved. It is unclear, though, what is a good criterion for measuring this lack of improvement. However, often it is useful to represent the collection of best trees so far in a compact way to allow scientists to monitor the reconstruction progress. Consensus and agreement trees are common such representations. Using existing static algorithms to produce these trees increases an already lengthy computational time substantially. In this paper we pr...

• Type: Conference
• Year: 2004
• Where: WABI
• Authors: Tanya Y. Berger-Wolf
• Keywords: Algorithms, Large-scale Phylogenetic Analyses, Optimal Tree, Phylogenetic Trees, WABI 2004