Home Models Trees Genomics PANDIT Jobs	Leaphy Likelihood estimation algorithms in phylogenetics Welcome to the homepage for the phylogenetic tree estimation software Leaphy. Leaphy uses maximum likelihood (ML) to estimate trees from aligned amino acid and nucleotide sequences under a variety of commonly used and popular models. The 1.0 version of Leaphy currently available is a beta release, which means that it is still in a state of development and will almost certainly contain errors, some of which may affect the performance of the program. These are more likely to cause problems in quirky datasets with extreme parameter values. If you use Leaphy in your research please cite: S. Whelan. New approaches to phylogenetic tree search and their application to large numbers of protein alignments. In press. Systematic Biology. If you would like a copy of the manuscript please contact me. Downloads      Leaphy binary for windows:  Leaphy.exe      Leaphy binary for linux: Leaphy      Leaphy binary for linux (64bit): Leaphy_64bit If you download the binaries please e-mail me (simon.whelan@manchester.ac.uk) so I can keep you informed of any important bug fixes and updates. Click here for links to all versions and a version history. Input data Leaphy currently supports data in the Phylip and sequential format. Click the links for example files for these  format. Making the program run Leaphy is intended to run from the command line, either under a shell for Linux, or under a Cygwin shell or the MS-DOS command prompt in Windows. The following instructions assume you are running under a shell; if you are using MS-DOS then drop the "./" before the program name. Running Leaphy using the interactive user interface 1. Type ./Leaphy to start the program 2. Input the name of a valid data file. Leaphy will determine whether the data contains amino acid or nucleotide sequences and if the file format is incorrect Leaphy will try and tell you why. 3. Choose a valid output file. All output from the program will be directed to this file. 4. Choose options using the interface and type "0" to start the analysis. Running Leaphy using only the command line ./Leaphy <data_file> <output_file> <Model> <Start_tree>   <data_file>: The input data file in either Phylip or sequential format. Leaphy will work out whether it is amino acid or nucleotide sequence data. <output_file>: The location of the output file. <Model>: Choose a starting model. DNA models include: JC, FEL, K2P, HKY, and REV. Amino acid models include: EQU, , JTT, DAY, mtREV, and WAG. These models usually have their equilibrium distribution of amino acids predefined. If you wish to use the observed frequencies from the data add "+F" to the model name. To add Г-distributed rate variation add "+dG" to the model. To add invariant sites add "+I" to the model. Example 1: "HKY+dG+I" would result in Leaphy running the HKY model with gamma distributed rates and invariant sites. The order of "+dG" and "+I" doesn't matter. Example 2: "DAY+F" would result in Leaphy running the dayhoff model with its equilibrium distribution of amino acids estimated from the sequence data. <Start_tree>: Where the tree estimation program starts. There are four options: "all": Try stepwise addition and bionj starting tree. "bionj": Only use a bionj starting tree. "SA": Only use a stepwise addition starting tree. "file=XXXX": Load an input tree from file XXXX. Tree files The tree files for Leaphy require a starting line with two numbers before the tree specifying the number of sequences in the tree and the number of trees. The tree is assumed to be in Newick format. For an example, click here. Reading results Leaphy outputs a file detailing the tree search. The optimal tree located by the program is at the bottom of this file in the line starting with the phrase "BestTree". The likelihoods calculated by Leaphy should not be used for comparing trees between programs and for comparing trees between the different runs of Leaphy. This is because approximations to the evolutionary model are made during the program's run, which may be different every time. I recommend using the programs in Ziheng Yang's PAML package for comparing likelihoods between trees. An example run Using the interactive menu 1. Download an example data file in sequential format. 2. Download a Leaphy executable (see Downloads section above). 3. Make sure both files are in the same directory 4. For windows MS-DOS type: "Leaphy"; For a linux or Cygwin shell type: "./Leaphy" 5. Type "data.seq" to input the data 6. Press return for the output file. 7. Type "0" to run Leaphy under the default "REV+dG" model. The same run from the command line would be: "./Leaphy data.seq data.seq.output REV+dG all" Both should give you an output file called data.seq.output. The final tree and its likelihood is located on the line towards the end of the file starting "BestTree".
Last modified: 11 June, 2007