MrBayes

MrBayes is a program for Bayesian inference and model choice across a wide range of phylogenetic and evolutionary models. MrBayes uses Markov chain Monte Carlo (MCMC) methods to estimate the posterior distribution of model parameters.

Finding available modules

[name@server ~]$ module spider mrbayes

Examples

Sequential

The following job script uses only one CPU core (--cpus-per-task=1). The example uses an input file (primates.nex) distributed with MrBayes.

File : submit-mrbayes-seq.sh

#!/bin/bash
#SBATCH --account=def-someuser  # replace with your PI account
#SBATCH --cpus-per-task=1 
#SBATCH --mem-per-cpu=3G        # increase as needed
#SBATCH --time=1:00:00          # increase as needed

module load mrbayes/3.2.7
cd $SCRATCH 
cp -v $EBROOTMRBAYES/share/examples/mrbayes/primates.nex .

mb primates.nex

The job script can be submitted with

[name@server ~]$ sbatch submit-mrbayes-seq.sh

Parallel

MrBayes can be run on multiple cores, on multiple nodes, and on GPUs.

MPI

The following job script will use 8 CPU cores in total, on one or more nodes. Like the previous example it uses an input file (primates.nex) distributed with MrBayes.

File : submit-mrbayes-parallel.sh

#!/bin/bash
#SBATCH --account=def-someuser  # replace with your PI account
#SBATCH --ntasks=8 				# increase as needed
#SBATCH --mem-per-cpu=3G        # increase as needed
#SBATCH --time=1:00:00          # increase as needed

module load mrbayes/3.2.7
cd $SCRATCH 
cp -v $EBROOTMRBAYES/share/examples/mrbayes/primates.nex .

srun mb primates.nex

The job script can be submitted with

[name@server ~]$ sbatch submit-mrbayes-parallel.sh

GPU

The following job script will use a GPU. Like the previous examples it uses an input file (primates.nex) distributed with MrBayes.

File : submit-mrbayes-gpu.sh

#!/bin/bash
#SBATCH --account=def-someuser  # replace with your PI account
#SBATCH --cpus-per-task=1
#SBATCH --gpus=1
#SBATCH --mem-per-cpu=3G        # increase as needed
#SBATCH --time=1:00:00          # increase as needed

module load gcc cuda/12 mrbayes/3.2.7
cd $SCRATCH 
cp -v $EBROOTMRBAYES/share/examples/mrbayes/primates.nex .

srun mb primates.nex

The job script can be submitted with

[name@server ~]$ sbatch submit-mrbayes-gpu.sh

Checkpointing

For users needing very long runs of MrBayes, we suggest you break up the work into several small jobs rather than one very long job. Long jobs have are more likely to be interrupted a hardware failure or a maintenance outage. Fortunately, MrBayes has a built-in mechanism for creating checkpoints, where progress can be saved from one job and continued in a subsequent job.

1. Create the first script (job).

File : job1.nex

execute primates.nex;

mcmc ngen=10000000 nruns=2 temp=0.02 mcmcdiag=yes samplefreq=1000 
stoprule=yes stopval=0.005 relburnin=yes burninfrac=0.1 printfreq=1000 
checkfreq=1000;

2. Create a second script (job).

File : job2.nex

execute primates.nex;

mcmc ngen=20000000 nruns=2 temp=0.02 mcmcdiag=yes samplefreq=1000
stoprule=yes stopval=0.005 relburnin=yes burninfrac=0.1 printfreq=1000
append=yes checkfreq=1000;

3. Create the submission script to run the smaller jobs

File : submit-mrbayes-cp.sh

#!/bin/bash
#SBATCH --account=def-someuser  # replace with your PI account
#SBATCH --ntasks=8 				# increase as needed
#SBATCH --mem-per-cpu=3G        # increase as needed
#SBATCH --time=1:00:00          # increase as needed
#SBATCH --array=1-2%1           # match the number of sub-jobs, only 1 at a time

module load gcc mrbayes/3.2.7

cd $SCRATCH 

# Copy one of the example locally on the local storage
cp -v $EBROOTMRBAYES/share/examples/mrbayes/primates.nex .

# Run using $SLURM_NTASKS
srun mb job${SLURM_ARRAY_TASK_ID}.nex

4. Submit the jobs

[name@server ~]$ sbatch submit-mrbayes-cp.sh

Message Passing Interface