Efficient Construction of the Evolutionary Tree from Quartets

Abstract

Scientists often need to use the information of the species to infer the evolutionary relationship among them.The evolutionary relationships are generally represented by a labeled binary tree, called the evolutionary tree(or phylogenetic tree). The problem of constructing the evolutionary tree for a set of species is often known as the phylogeny problem. The difficulty of such problem is that the number of possible evolutionary trees is very large. There are many methods for solving the phylogeny problem including parsimony, mazimum likelihood, divide-and-conquer approach and distance matrix methods[1,5,6,7,8].As the number of species increases, so does the number of possible trees,making it difficult to exhaustively enumerate all possible relationships. The quantitative nature of species relationships therefore requires the development of more rigorous methods for tree construction. in this paper, we proposed a new effective approach to construct the evolutionary tree. Our approach consists of three steps. First, we will partition the species and dispath them onto different computers in order to quickly find the candidate quartets. Second, each computer will examine its local candidate quqrtets for compatibility. Third, a procedure will be applied to merge and evaluate compatible quartets found on each computer to form even larger quartets, with which we can construct evolutionary trees.