A GA for maximum likelihood phylogenetic inference using neighbour-joining as a genotype to phenotype mapping

Evolutionary relationships among species can be represented by a phylogenetic tree and inferred by optimising some measure of fitness, such as the statistical likelihood of the tree (given a model of the evolutionary process and a data set). The combinatorial complexity of inferring the topology of the best tree makes phylogenetic inference ideal for genetic algorithms. In this paper, two existing algorithms for phylogenetic inference (neighbour-joining and maximum likelihood) are co-utilised within a GA and enable the phenotype and genotype to be assigned quite different representations. The exploration vs. exploitation aspects of the algorithm are examined in some test cases. The GA is compared to the well known phylogenetic inference program PHYLIP. Categories and Subject Descriptors J.3 [Life and medical sciences]: Biology and genetics; F.2.2 [Nonnumerical algorithms and problems]: Sorting and searching; I.5.3 [Pattern recognition]: Clustering General Terms algorithms Keywords p...

• Type: Conference
• Year: 2005
• Where: GECCO
• Authors: Leon Poladian
• Keywords: GECCO 2005, Genetic Algorithms, Phylogenetic Inference, Phylogenetic Inference Ideal