Improving the efficiency of SPR moves in Phylogenetic Tree Search Algorithms based on Maximum-Likelihood.

Please cite:
"Improving the efficiency of SPR moves in Phylogenetic Tree Search Algorithms based on Maximum-Likelihood."
Hordijk W., Gascuel O.
Bioinformatics, 21 (24), 4338-4347, 2005.


Here you can find executables and data sets as used in our Bioinformatics paper. We would like to note, however, that the program is still very much in a development stage, and is not guaranteed to be bug-free, crash-free, or in any other way problem-free.


The SPR algorithm is implemented in PhyML, and uses the same arguments as the original PhyML 2.4.4 program. By default, the additional SPR parameters (N_RGRFT, N_GLOBL, and MAX_DIST) are set to 10%, 5%, and 5% of the number of taxa in the data set, respectively. However, if a file "params.spr" exists in the current working directory, the program will read the parameter values from this file. They should be given on one (the first) line, separated by one or more spaces, e.g.: 30 20 10 which will set the parameter values to N_RGRFT=30, N_GLOBL=20, and MAX_DIST=10, in that order.


Two binaries are available. The first one, 'spr', uses the SPR algorithm alone. The second one, 'ph_spr', uses a combination of the original PhyML and new SPR algorithms. The programs were compiled on a linux system. Finally, the seven data sets used in our paper are also available here.

Binaries (spr and ph_spr) are available here.

Data sets are available here.


For any questions about the algorithm, the options of the program or the source code, you can contact: