To install RAxML on the cluster and be a part of your profile so you don't have to load it every time:
vi ~/.bash_profile
i to insert the following lines:
module load openmpi
module load gcc/10.2.0
export PATH=/project2/cbrook/software/raxml-ng/bin:$PATH
So it should look like this:
# .bash_profile
# Get the aliases and functions
if [ -f ~/.bashrc ]; then
. ~/.bashrc
fi
# User specific environment and startup programs
PATH=$PATH:$HOME/bin
export PATH
module load openmpi
module load gcc/10.2.0
export PATH=/project2/cbrook/software/raxml-ng/bin:$PATH
~
~
~
~
~
~
~
~
"~/.bash_profile" 15L, 276C 1,1 All
press esc to edit insert mode, then type :wq to save and quit
check it works by requesting version (if it says ‘not found’, didn’t work): raxml-ng-mpi --version
Below is an example of a SLURM script I ran for a very large (412 genera) picornavirus phylogeny, the max time alloted by the cluster (36 hours) was not enough to finish so I just emailed the cluster to extend the job time to around 96 hours just to be safe:
#!/bin/bash
#SBATCH --job-name=refseq_all_picorna_raxml
#SBATCH --partition=broadwl
#SBATCH --output=refseq_all_picorna_raxml.out
#SBATCH --nodes=1
#SBATCH --ntasks=
#SBATCH --ntasks-per-node=
#SBATCH --time=36:00:00
module load vim/7.4
module load emacs/25.1
module load python/3.6
module load java/1.8.0_121
module load cmake/3.15.1
raxml-ng-mpi --all --msa RAxML_aligned_all_refseq_picornaviruses.fasta --model GTR+G4 --prefix T18 --seed 6 --threads 10 --bs-metric fbp,tbe
Note that you need to have matching numbers for the -threads command in RAxML with the --ntasks command call to SLURM
Great troubleshooting guide: https://github-wiki-see.page/m/amkozlov/raxml-ng/wiki/Tutorial Note: if raxml-ng stops unexpectedly or does not converge, you may need to manually change internal cutoff value or number of bootstraps. info on how to do so in the link above