It is well known that the use of different probabilistic models of evolution may change the results of phylogenetic analyses, and during the last years much research has been devoted to the selection of best-fit models. An alternative to model selection is model averaging or multimodel inference, where all models are simultaneously used to infer different parameters of the substitution process, or the phylogenetic tree (topology and branches) itself. Here I will discuss some preliminary simulations that suggest that model-averaging might not be very useful to estimate substitution parameters, but that might reduce the variance associated to the estimation of tree topology. New software that implements model averaging for phylogenetics will also be introduced.
Related Links
* http://darwin.uvigo.es - David Posada's web site
view more