Study And Improvement Of Methods Of Constructing Phylogenetic Tree

An evolutionary tree，also called Phylogenetic tree，is a tree-like structure that willdescribe the genetic relationship between organisms. Reconstructing the evolutionaryhistory of a set of species is an elementary problem in biology. An evolutionary treecontaining all the species is known as the tree of life, the tree describes the relationshipof all the species. It is very meaningful to reconstruct a high-quality evolutionary tree. Itcan help us to understand the evolutionary history and helps to inference themechanisms. And an evolutionary tree is also helpful for other fields, such as moleculargenetics, biomedicine, gene discovery, well-bred breeding, etc.Currently, the methods for reconstructing an evolutionary tree can be divided intotwo categories: distance-based methods and sequence-based methods. In this paper, wewill focus on the distance methods, especially neighbor-joining methods.First, propose a new distance method to solve the not-estimated value problem. In adistance matrix, because of the short sequence length, evolutionary model applicability,distance formula feasibility, and other problems, some unknown values will occur. Inthis paper we propose a new combined distance. Second, propose an improve method ofneighbor-joining method to solve the tied tree problem and improve the performance ofthe original method. At last, we develop an evolutionary tree reconstructing system.The system combines all the work above.The results show that the new distance is better in statistical and practice. Thestatistical values show that the new distance compared with the original one hasdifferent degree of improvement. The results of the new method and the classical onehave identical topology, if containing unknown values, our method performs better. Andthe improved method runs faster the neighbor-joining method. In most cases, theimproved method will return only one tree, and don’t rely on the gene order.