Scalable Phylogenetics through Input Preprocessing

Phylogenetic reconstruction is one of the fundamental problems in computational biology. The combinatorial explosion of the state space and the complexity of mathematical models impose practical limits on workable problem sizes. In this article we explore the scalability of popular algorithms under real datasets as problem dimensions grow. We furthermore develop an efficient preclassification and partitioning strategy based on guide trees, which are used to intently define an evolutionary hierarchy of groups of related data, and to determine membership of individual data to their corresponding subproblems. Finally, we apply this method to efficiently calculate exhaustive phylogenies of human mitochondrial DNA according to phylogeographic criteria.