Version

RAxML  7.0.4 and 7.2.8 installed on Hippu.

RAxML  7.0.4 and 7.2.8 installed on Murska.

RAxML  7.0.4 and 7.2.8 installed on Vuori.

Description

RAxML is a fast program for the inference of phylogenies with maximum likelihood method. RAxML offers several different evolutionary models for both DNA and amino acid sequences.

Usage

RAxML needs to be initialized to be used. On all systems (Hippu, Murska, Vuori) default version is loaded with command

module load raxml

To load a specific version you can give the version number as shown by command "module avail", e.g.

module load raxml/7.2.8

On Hippu only the serial version (raxmlHPC) is installed. On Murska and Vuori also parallel version (raxmlHPC-MPI) is available.

 The basic call (depending on the version) for DNA sequence data (uses GTR + 4 gamma categories):

raxmlHPC -n run_name -s alignment -m GTRGAMMA

and for amino acid sequences (uses JTT + 4 gamma categories):

raxmlHPC -n run_name -s alignment -m PROTGAMMAJTT

RAxML offers special evolutionary models, where alignment columns are grouped in a number of similarly behaving (in respect to substitution frequencies) columns. This is very close to usual gamma models, but is much faster to compute. Model applying this method are marked with "CAT". Usual gamma models are marked with "GAMMA", and plain models are called with their simple name, like "HKY85".

More information about the RAxML-HPC version is available in its manual.

Batch jobs in Murska

Running RAxML analysis might take considerable amount of CPU time and memory. To submit a long RAxML run, you must construct a command file that includes queue options and setup and execution commands for RAxML.

Notice that the Murska computing nodes can not see $HOME directory, so files must be copied to $WRKDIR to run a job.

An example batch file for running a serial job on Murska:

#!/bin/csh

## name of your job, %J will show as your jobID
#BSUB -J raxml_%J

## system error message output file
#BSUB -e raxml_err_%J

## system message output file
#BSUB -o raxml_out_%J

## a per-process (soft) memory limit
## limit is specified in KB
## example: 1 GB is 1048576
#BSUB -M 4194304

## how long a job takes, wallclock time hh:mm
#BSUB -W 04:00

## number of proceses
#BSUB -n 1

## You can give the module load command before submitting the job or include it here
module load raxml

# Modify this RAxML command to your liking
srun raxmlHPC -n test1 -s example.phy -m GTRGAMMAI

An example batch file for running a parallel job on Murska:

#!/bin/csh

## name of your job, %J will show as your jobID
#BSUB -J raxml_%J

## system error message output file
#BSUB -e raxml_err_%J


## system message output file
#BSUB -o raxml_out_%J

## a per-process (soft) memory limit
## limit is specified in KB
## example: 1 GB is 1048576
#BSUB -M 2097152

## how long a job takes, wallclock time hh:mm
#BSUB -W 04:00

## number of proceses
#BSUB -n 8

## You can give the module load command before submitting the job or include it here
module load raxml

# Modify this RAxML command to your liking
mpirun -srun raxmlHPC -n test1 -s example.phy -m GTRGAMMAI

The batch job is submitted with command (where "my_job_script" is the name of your batch file):

bsub < my_job_script

Batch jobs in Vuori

Running RAxML analysis might take considerable amount of CPU time and memory. To submit a long RAxML run, you must construct a command file that includes queue options and setup and execution commands for RAxML.

Notice that the Vuori computing nodes can not see $HOME directory, so files must be copied to $WRKDIR to run a job.

You can modify the following example batch file accordingly where indicated:

#!/bin/csh

## Name of job
#SBATCH -J my_jobname

## system error message output file
#SBATCH -e my_output_err_%j

## system message output file
#SBATCH -o my_output_%j

## a per-process (soft) memory limit
## limit is specified in MB
## example: 1 GB is 1000
#SBATCH --mem-per-cpu=1000

## how long a job takes, wallclock time hh:mm:ss
#SBATCH -t 10:00:00

##the number of processes (number of cores)
#SBATCH -n 12

## On Vuori there is 12 cores per node, so there could be 12 mpi-processes per node
## This just minimizes the number of nodes in your job.
#SBATCH --ntasks-per-node=12

##parallel queue 
#SBATCH -p parallel

## run my MPI executable
## Modify the command line as necessary
srun raxmlHPC-MPI -N 100 -n test1 -s cox1.phy -m GTRGAMMAI

The batch job is submitted with command (where "my_job_script" is the name of your batch file):

sbatch my_job_script

Manual

Manual can be downloaded from the developer site.

You might also want to read the worked-out examples in Molecular systematics - material for practicals.

Distribution

RAxML can be downloaded from it's website. In addition to the source code, DOS executables are available.

User support