Les méthodes probabilistes en phylogénie moléculaire : (2) L'approche bayésienne

Frédéric Delsuc (1) , Emmanuel J.P. Douzery (1)

1 UMR ISEM - Institut des Sciences de l'Evolution de Montpellier

Although being of relatively recent introduction into the field, the Bayesian approach of phylogenetic reconstruction has soon imposed itself as a leading method. This probabilistic method is based on the calculation of posterior probabilities of phylogenetic trees by the combination of a prior probability with the likelihood function. The use of Monte Carlo Markov chains (MCMC) to compute Bayesian posterior probabilities makes this approach particularly appealing. Indeed, these numerical methods allow the implementation of complex models of sequence evolution incorporating a large number of parameters in a reasonable computation time. The Bayesian approach thus appears as articularly promising for the future of phylogenetics. However, as a recently developed method, this approach has its still unresolved problems and the evaluation of its properties is still in its infancy. This latest aspect thus represents a major research axis for the next coming years.

Domaines: Evolution; Biodiversité; Biosystématique; Bio-informatique [q-bio.QM]; Bio-Informatique, Biologie Systémique [q-bio.QM]

Frédéric Delsuc, Emmanuel J.P. Douzery. Les méthodes probabilistes en phylogénie moléculaire : (2) L'approche bayésienne. Biosystema, 2004, 22, pp.75-86. ⟨halsde-00193038⟩