"RAxML is a fast implementation of maximum-likelihood (ML) phylogeny estimation that operates on both nucleotide and protein sequence alignments."
Availability and Restrictions
The following versions of RAxML are available on OSC systems:
You can use
module spider raxml to view available modules for a given machine. Feel free to contact OSC Help if you need other versions for your work.
RAxML is available to all OSC users without restriction.
Publisher/Vendor/Repository and License Type
The Exelixis Lab, Open source
On the Oakley Cluster RAxML is accessed by executing the following commands:
module load raxml
raxmlHPC[-MPI|-PTHREADS] -s sequenceFileName -n outputFileName -m substitutionModel [-a weightFileName] [-b bootstrapRandomNumberSeed] [-c numberOfCategories][-d] [-e likelihoodEpsilon] [-E excludeFileName] [-f a|b|c|d|e|g|h|i|j|m|n|o|p|s|t|w|x] [-g groupingFileName] [-h] [-i initialRearrangementSetting] [-j] [-k] [-l sequenceSimilarityThreshold] [-L sequenceSimilarityThreshold] [-M] [-o outGroupName1[,outGroupName2[,...]]] [-p parsimonyRandomSeed] [-P proteinModel] [-q multipleModelFileName] [-r binaryConstraintTree] [-t userStartingTree] [-T numberOfThreads] [-u multiBootstrapSearches] [-v][-w workingDirectory] [-x rapidBootstrapRandomNumberSeed][-y][-z multipleTreesFile] [-#|-N numberOfRuns]
-aSpecify a column weight file name to assign individual weights to each column of the alignment. Those weights must be integers separated by any type and number of whitespaces within a separate file, see file "example_weights" for an example.
-bSpecify an integer number (random seed) and turn on bootstrapping. DEFAULT: OFF
-cSpecify number of distinct rate catgories for RAxML when modelOfEvolution is set to GTRCAT or GTRMIX. Individual per-site rates are categorized into numberOfCategories rate categories to accelerate computations. DEFAULT: 25
-dstart ML optimization from random starting tree. DEFAULT: OFF
-eset model optimization precision in log likelihood units for final optimization of tree topology under MIX/MIXI or GAMMA/GAMMAI. DEFAULT: 0.1 for models not using proportion of invariant sites estimate 0.001 for models using proportion of invariant sites estimate
-Especify an exclude file name, that contains a specification of alignment positions you wish to exclude. Format is similar to Nexus, the file shall contain entries like "100-200 300-400", to exclude a single column write, e.g., "100-100", if you use a mixed model, an appropriately adapted model file will be written.
-fselect algorithm: "
-f a": rapid Bootstrap analysis and search for best-scoring ML tree in one program run "
-f b": draw bipartition information on a tree provided with "-t" based on multiple trees (e.g. form a bootstrap) in a file specifed by "-z" "
-f c": check if the alignment can be properly read by RAxML "
-f d": new rapid hill-climbing "
-f e": optimize model+branch lengths for given input tree under GAMMA/GAMMAI only "
-f g": compute per site log Likelihoods for one ore more trees passed via "-z" and write them to a file that can be read by CONSEL "
-f h": compute log likelihood test (SH-test) between best tree passed via "-t" and a bunch of other trees passed via "-z" "
-f i": perform a really thorough bootstrap, refinement of final BS tree under GAMMA and a more exhaustive algorithm "
-f j": generate a bunch of bootstrapped alignment files from an original alignment file "
-f m": Compare bipartitions between two bunches of trees passed via "-t" and "-z" respectively. This will return the Pearson correlation between all bipartitions found in the two tree files. A file called RAxML_bipartitionFrequencies.outpuFileName will be printed that contains the pair-wise bipartition frequencies of the two sets "
-f n": Compute the log likelihood score of all trees contained in a tree file provided by "-z" under GAMMA or GAMMA+P-Invar "
-f o": old and slower rapid hill-climbing "
-f p": perform pure stepwise MP addition of new sequences to an incomplete starting tree "
-f s": split up a multi-gene partitioned alignment into the respective subalignments "
-f t": do randomized tree searches on one fixed starting tree "
-f w": compute ELW test on a bunch of trees passed via "-z" "
-f x": compute pair-wise ML distances, ML model parameters will be estimated on an MP starting tree or a user-defined tree passed via "-t", only allowed for GAMMA-based models of rate heterogeneity. DEFAULT: new rapid hill climbing
-gspecify the file name of a multifurcating constraint tree this tree does not need to be comprehensive, i.e. must not contain all taxa
-hDisplay this help message.
-iInitial rearrangement setting for the subsequent application of topological changes phase. DEFAULT: determined by program
-jSpecifies if checkpoints will be written by the program. If checkpoints (intermediate tree topologies) shall be written by the program specify "-j" DEFAULT: OFF
-kSpecifies that bootstrapped trees should be printed with branch lengths. The bootstraps will run a bit longer, because model parameters will be optimized at the end of each run. Use with CATMIX/PROTMIX or GAMMA/GAMMAI. DEFAULT: OFF
-lSpecify a threshold for sequence similarity clustering. RAxML will then print out an alignment to a file called sequenceFileName.reducedBy.threshold that only contains sequences <= the specified thresold that must be between 0.0 and 1.0. RAxML uses the QT-clustering algorithm to perform this task. In addition, a file called RAxML_reducedList.outputFileName will be written that contains clustering information. DEFAULT: OFF
-LSame functionality as "-l" above, but uses a less exhasutive and thus faster clustering algorithm. This is intended for very large datasets with more than 20,000-30,000 sequences. DEFAULT: OFF
-mModel of Nucleotide or Amino Acid Substitution: NUCLEOTIDES: "
-m GTRCAT" : GTR + Optimization of substitution rates + Optimization of site-specific evolutionary rates which are categorized into numberOfCategories distinct rate categories for greater computational efficiency if you do a multiple analysis with "-#" or "-N" but without bootstrapping the program will use GTRMIX instead "
-m GTRGAMMA" : GTR + Optimization of substitution rates + GAMMA model of rate heterogeneity (alpha parameter will be estimated) "
-m GTRMIX" : Inference of the tree under GTRCAT and thereafter evaluation of the final tree topology under GTRGAMMA "
-m GTRCAT_GAMMA" : Inference of the tree with site-specific evolutionary rates. However, here rates are categorized using the 4 discrete GAMMA rates. Evaluation of the final tree topology under GTRGAMMA "
-m GTRGAMMAI" : Same as GTRGAMMA, but with estimate of proportion of invariable sites "
-m GTRMIXI" : Same as GTRMIX, but with estimate of proportion of invariable sites "
-m GTRCAT_GAMMAI" : Same as GTRCAT_GAMMA, but with estimate of proportion of invariable sites AMINO ACIDS: "
-m PROTCATmatrixName[F]" : specified AA matrix + Optimization of substitution rates + Optimization of site-specific evolutionary rates which are categorized into numberOfCategories distinct rate categories for greater computational efficiency if you do a multiple analysis with "-#" or "-N" but without bootstrapping the program will use PROTMIX... instead "
-m PROTGAMMAmatrixName[F]" : specified AA matrix + Optimization of substitution rates + GAMMA model of rate heterogeneity (alpha parameter will be estimated) "
-m PROTMIXmatrixName[F]" : Inference of the tree under specified AA matrix + CAT and thereafter evaluation of the final tree topology under specified AA matrix + GAMMA "
-m PROTCAT_GAMMAmatrixName[F]" : Inference of the tree under specified AA matrix and site-specific evolutionary rates. However, here rates are categorized using the 4 discrete GAMMA rates. Evaluation of the final tree topology under specified AA matrix + GAMMA "
-m PROTGAMMAImatrixName[F]" : Same as PROTGAMMAmatrixName[F], but with estimate of proportion of invariable sites "
-m PROTMIXImatrixName[F]" : Same as PROTMIXmatrixName[F], but with estimate of proportion of invariable sites "
-m PROTCAT_GAMMAImatrixName[F]" : Same as PROTCAT_GAMMAmatrixName[F], but with estimate of proportion of invariable sites Available AA substitution models: DAYHOFF, DCMUT, JTT, MTREV, WAG, RTREV, CPREV, VT, BLOSUM62, MTMAM, GTR With the optional "F" appendix you can specify if you want to use empirical base frequencies. Please not that for mixed models you can in addition specify the per-gene AA model in the mixed model file (see manual for details)
-MSwitch on estimation of individual per-partition branch lengths. Only has effect when used in combination with "-q". Branch lengths for individual partitions will be printed to separate files. A weighted average of the branch lengths is computed by using the respective partition lengths. DEFAULT: OFF
-nSpecifies the name of the output file.
-oSpecify the name of a single outgrpoup or a comma-separated list of outgroups, eg "-o Rat" or "-o Rat,Mouse", in case that multiple outgroups are not monophyletic the first name in the list will be selected as outgroup, don't leave spaces between taxon names!
-qSpecify the file name which contains the assignment of models to alignment partitions for multiple models of substitution. For the syntax of this file please consult the manual.
-pSpecify a random number seed for the parsimony inferences. This allows you to reproduce your results and will help me debug the program. This option HAS NO EFFECT in the parallel MPI version
-PSpecify the file name of a user-defined AA (Protein) substitution model. This file must contain 420 entries, the first 400 being the AA substitution rates (this must be a symmetric matrix) and the last 20 are the empirical base frequencies
-rSpecify the file name of a binary constraint tree. This tree does not need to be comprehensive, i.e. must not contain all taxa
-sSpecify the name of the alignment data file in PHYLIP format
-tSpecify a user starting tree file name in Newick format
-TPTHREADS VERSION ONLY! Specify the number of threads you want to run. Make sure to set "-T" to at most the number of CPUs you have on your machine, otherwise, there will be a huge performance decrease!
-uSpecify the number of multiple BS searches per replicate to obtain better ML trees for each replicate. DEFAULT: One ML search per BS replicate
-vDisplay version information
-wName of the working directory where RAxML will write its output files. DEFAULT: current directory
-xSpecify an integer number (random seed) and turn on rapid bootstrapping
-yIf you want to only compute a parsimony starting tree with RAxML specify "-y", the program will exit after computation of the starting tree. DEFAULT: OFF
-zSpecify the file name of a file containing multiple trees e.g. from a bootstrap that shall be used to draw bipartition values onto a tree provided with "-t", It can also be used to compute per site log likelihoods in combination with "-f g" and to read a bunch of trees for a couple of other options ("-f h", "-f m", "-f n").
-#|-NSpecify the number of alternative runs on distinct starting trees. In combination with the "-b" option, this will invoke a multiple bootstrap analysis. Note that "-N" has been added as an alternative since "-#" sometimes caused problems with certain MPI job submission systems, since "-#" is often used to start comments. DEFAULT: 1 single analysis
The file brown.philip is from the PAML example and modified to be in PHYLIP format.
#PBS -N RAxML_test #PBS -l walltime=10:00 #PBS -l nodes=1:ppn=4 cd $PBS_O_WORKDIR module load raxml raxmlHPC-PTHREAD -T 4 -m GTRCAT -n test -s brown.phylip
If you want to compute more than one tree add the following line before calling the program:
module load pvfs2
- RAxML homepage
- By typing raxmlHPC-PTHREAD -help after you have loaded the two modules.