Supertree Methods for Building the Tree of Life: Divide-and-Conquer Approaches to Large Phylogenetic Problems
Reconstructing the tree of life will require fast methods for building very large phylogenetic trees from patchy data. The leading candidates for such an approach employ supertree methods as part of a divide-and-conquer strategy. Here, we discuss two aspects of a phylogenetic divide-and-conquer method: the decomposition of the tree into subproblems and the recombining of these into an overall solution. In particular, we highlight and explore the issue of effective taxon overlap, how it might be achieved via suitable decomposition and how it might be used to guide the setting of priorities for additional data acquisition, and we show how some knowledge of phylogeny is vital in both contexts. Last, we show that quartet puzzling, the best known phylogenetic divide-andconquer method, can perform poorly when not all quartets are available, and we present a new fast supertree method designed to perform better in this context. Whilst a great deal of work remains, such an approach has great potential as part of a divide-and-conquer method for reconstructing large phylogenies on the scale of the tree of life or for large subsets of species rich taxa.