MrBayes: Bayesian inference of phylogenies
MrBayes is an extremely popular program for the Bayesian inference of phylogenies based on molecular sequence data and/or character state data. It is included in the Rocks Bio Roll and is therefore available on all HPC nodes.
1. Upload necessary files onto the HPC
More information on file management,
connecting to the hpc and ssh can be found here.
First create a directory called mb inside your home directory. Upload your nexus file to
that directory. For this example I will use an example file included with Mr. Bayes, bglobin.nex. I changed one of the
last lines of that file from mcmcp to mcmc, in order to have some output.
After the upload your should have one file inside your directory.
2. Create Submission script.
More information on MOAB scripts can be found here.
More information on editing text files can be found here
Create a text called dnaml.moab.sh in your phylip directory. The contents of this file should look like this.
# Name this job.
#MOAB -N dnaml
# Run this job in the "backfill" queue. Delete the following line if
# you expect this to run more than 4 hours.
#MOAB -q backfill
# Maximum wallclock time :( Hours : Minutes : Seconds )
# Change this to the maximum length of time you wish your program
# to run
#MOAB -l walltime=03:00:00
# Join the two output files (standard output and standard error)
# into one. Based on the job name, this means the single output file
# will be called "dnaml.oJOBID" where JOBID is a job id number
# assigned when the job is submitted.
#MOAB -j oe
# This command moves from your home directory to the directory
# from which this script was submitted.
# First you have to source the script that sets up the Bio Applications
# The command to run
3. Submit the job to run on the hpc.
More information on submitting jobs and different MOAB commands can be found here
In order to submit the job to be run on the cluster, use the program msub.
4. Monitoring the job.
More information on monitoring jobs and different MOAB commands can be found here
In order to see the status of all your jobs at the command line type.
showq -w user=`whoami`
5. Collect the output.
Once the job is funished running there will be files bglobin.nex.mcmc bglobin.nex.run1.p bglobin.nex.mcmc.run1.t bglobin.nex.mcmc.run2.t bglobin.nex.mcmc.run2.p inside the mb directory. Transfer these back to your computer to collect the output.