Construction of Phylogenetic Trees

Abstract

A phylogenetic tree is an unordered, distinctly leaf-labeled tree which shows evolutionary relationship among the leaves that represent the species. Inferring phylogenies using computational methods has several important applications in biological and biochemical research. Drug discovery and conservation biology are such examples. Phylogenetic trees have some important applications in practical fields, such as forensics, gene and protein function detection and drug design. In Bioinformatics, computation of the Tree of Life considering all living beings on earth is a grand challenge, in which phylogenetic tree can play important role to derive such relations. Phylogenetic tree construction from valid triplet set is one of the most common approaches of construction till date. A valid triplet set consists of a number of rooted triplets which are individually a phylogenetic tree of three leaves. Aho et al: first investigate the problem and later other researchers gradually meliorate and enrich his idea. These great research works solve many problems as well as introduce some open problems related to phylogenetic tree construction. Such a problem is to construct a phylogenetic tree consistent with all triplets in the given triplet set. Experimentally obtained data may contain incorrect information which results in erroneous triplet set, and thus maximum rooted triplet consistency (MaxRTC) is maintained in this case. During phylogenetic tree construction, maintenance of MaxRTC, sometimes ignores some important triplets where loss of information occurs. Moreover, the tree constructed from the triplet set contains unevenly distributed species. In this thesis, we study different aspects of the problem and introduce a heuristic algorithm to construct phylogenetic tree on all the triplets in the triplet set by making a small number of corrections on the triplets, where approximately even distribution of species is ensured. For a given set of m triplets with n species, this algorithm runs in O(m n) time. Though the algorithm is less efficient, the loss of information is less here than previous algorithms. On the other hand, checking the consistency of the triplet in a phylogenetic tree is another important problem in the field of phylogenetics. It is useful to compare two phylogenetic trees and measure the dissimilarity between them to check the accuracy. So it is very important to check this consistency in minimum amount of time. In this thesis, we give an algorithm to check the consistency of the triplet in constant time. This algorithm requires an initialization stage of time complexity O(n2) that preprocess the phylogenetic tree. Then we check if the triplet satisfies the phylogenetic tree or not in O(1) time. Though the initial complexity is high, the algorithm is helpful because it can check the consistency of a triplet in constant time.