Protein phylogenetic inference using maximum likelihood with a genetic algorithm.

This paper presents a method to construct phylogenetic trees from amino acid sequences. Our method uses a maximum likelihood approach which gives a confidence score for each possible alternative tree. Based on this approach, we developed a genetic algorithm for exploring the best score tree: randomly generated alternative trees are rearranged so that their scores are improved by utilizing crossover and mutation operators. In a test of our algorithm on a data set of EF-1 alpha sequences, we found that the performance of our algorithm is comparable to that of other tree-construction methods.