Bioinformatics for DNA Sequence Analysis.pdf - Index of

Chapter 5Selection of Models of DNA Evolution with jMODELTESTDavid PosadaAbstractjMODELTEST is a bioinformatic tool for choosing among different models of nucleotide substitution.The program implements five different model selection strategies, including hierarchical and dynamicallikelihood ratio tests (hLRT and dLRT), Akaike and Bayesian information criteria (AIC andBIC), and a performance-based decision theory method (DT). The output includes estimates ofmodel selection uncertainty, parameter importances, and model-averaged parameter estimates,including model-averaged phylogenies. jMODELTEST is a Java program that runs under Mac OSX,Windows, and Unix systems with a Java Run Environment installed, and it can be freely downloadedfrom http://darwin.uvigo.es.Key words: Model selection, likelihood ratio tests, AIC, BIC, performance-based selection, statisticalphylogenetics.1. IntroductionPhylogenetic reconstruction from DNA sequences is a problemof statistical inference. Since statistical inferences cannotbe drawn in the absence of probabilities, the use of models ofnucleotide substitution to calculate probabilities of changebetween nucleotides along the branches of a phylogenetictree is essential for many evolutionary and comparative analyses.Importantly, the use of different models of DNA evolutioncan change the outcome of the phylogenetic analysis. Theparameters affected include branch lengths, transition/transversionratio, overall divergence, and rate variation amongsites, whose estimates can be biased under a wrong modelDavid Posada (ed.), Bioinformatics for DNA Sequence Analysis, Methods in Molecular Biology 537ª Humana Press, a part of Springer ScienceþBusiness Media, LLC 2009DOI 10.1007/978-1-59745-251-9_593