PhyML is a software that estimates maximum likelihood phylogenies from alignments of nucleotide or amino acid sequences. The main strength of PhyML lies in the large number of substitution models coupled to various options to search the space of phylogenetic tree topologies, going from very fast and efficient methods to slower but generally more accurate approaches. PhyML was designed to process moderate to large data sets. In theory, alignments with up to 4,000 sequences 2,000,000 character-long can be processed.

NEWS!

// 12/04/2012. Improved versions of the NNI and SPR search algorithms are now available in PhyML. In particular, NNI shows much improved performance compared to previous releases. As usual, any feedback is welcome.

// 08/03/2012. PhyML implements a covarion model. It is not documented yet but has been tested thoroughly by Salvador Capella. Please feel free to send me an email (s.guindon at auckland.ac.nz) for more information about setting up a PhyML analysis using this model.

// 08/03/2012. The tree topology constraint feature of PhyML went through a few tests (thanks to Jaime Huerta Cepas) and is now more stable. It is also documented in PhyML user manual (see download on the left).

// 09/02/2012. PhyML implements tree topology estimation under user-defined clade constraints. This option is not documented yet and is only available as a beta version. For people interested in testing it, please feel free to send an email to s.guindon at auckland.ac.nz for more information.

// 09/06/2011. PhyML now has a discussion group: http://groups.google.com/group/phyml-forum

// 08/03/2011. PhyML now implements a new mixture model of rate variation across sites where the relative rates are not constrained to be distributed according to a gamma distribution. The rate in each category of the mixture is `freely' estimated from the data. While involving more parameters than the gamma model (2c-1 versus 1, where c is the number of rate classes), this model often provides a better fit to the data. Thanks to Julien Soubrier for testing this option.