Uma metaheurística para tratar a parcimônia em árvores filogenéticas

Abstract

In society, there are several products with genetically altered resources, from food to medical treatments. Whether to remove or add features in a given species, researchers need to know the degrees of relatedness between the species under study. To understand the kinship relationships between species, it is necessary to represent them in the phylogenetic tree, according to their characteristics. To build a phylogenetic tree, similarity and distance methods are used, both of which have performance problems or conflicting results. In the method by distance, the Principle of Parsimony is used, it is stated that the ideal tree would have the least amount of modifications. The phylogenetic tree construction algorithms present parsimony problems when there are more than one species with the shortest distance, which limits the search space for descendants or ancestors of close species. A metaheuristic is then suggested to solve this impasse, after modifications, to define that classifying the species of three types of parsimony by number of times selected together with likelihood helps Wagner’s Algorithm to get out of maximum global ruins, resulting in a tree more realist.