Selection of Phylogenetic Models of Molecular Evolution

Abstract

Abstract The use of different models of molecular evolution can change the conclusions derived from the evolutionary analysis of deoxyribonucleic acid (DNA) and protein sequence alignments. Several methods have been developed for the selection of the probabilistic model of nucleotide substitution or amino acid replacement that best fits the particular data at hand. Simulation studies indicate that these techniques work very well, and in recent years these methods have been implemented in a number of programs such as ModelTest, ProtTest and recently jModelTest. These programs also provide various tools for quantifying uncertainty in model selection, model averaging or multimodel inference. Key Concepts: Models of molecular evolution allow us to calculate probabilities of change between nucleotide and amino acid sequences. The use of different models of evolution can change the outcome of the phylogenetic analysis. Different datasets can be best‐fitted by distinct models. Model selection techniques are quite accurate at identifying the generating model in simulations. Programs like jModelTest and ProtTest facilitate the routinary selection of models of evolution in phylogenetics.