Minimizing Phylogenetic Number to find Good Evolutionary Trees

Inferring phylogenetic trees is a fundamental problem in computational biology. We present a new objective criterion, the phylogenetic number, for evaluating evolutionary trees for species deened by biomolecular sequences or other qualitative characters. The phylogenetic number of a tree T is the maximum number of times that any given character state arises in T. By contrast, the classical parsimony criterion measures the total number of times that diierent character states arise in T. We consider the following related problems: nding the tree with minimum phyloge-netic number, and computing the phylogenetic number of a given topology in which only the leaves are labeled by species. When the number of states is bounded (as is the case for biomolecular sequence characters), we can solve the second problem in polynomial time. Given the topology for an evolutionary tree, we can also compute a phylogeny with phylogenetic number 2 (when one exists) for an arbitrary number of states. This algorithm can be used to further distinguish trees that are equal under parsimony. We also consider a number of other related problems.