nextflow_setup.md

How to Install Nextflow on the Cluster Using Anaconda

1. Access one of the login nodes of the cluster.

2. Check if Anaconda is available on the cluster:

   module spider anaconda3

3. If Anaconda is available, load the module. The following is an example for SDSC Expanse:

   module purge
   module load cpu/0.15.4
   module load slurm
   module load anaconda3/2020.11
   eval "$(conda shell.bash hook)"

4. Set up Conda on the HPC cluster:

   • Create necessary directories and symlink for Conda packages to improve the speed of environment generation:

   mkdir /home/$USER/.conda
   mkdir -p /tmp/$USER/conda_pkgs
   ln -s /tmp/$USER/conda_pkgs /home/$USER/.conda/pkgs
   mkdir -p ~/a/conda_envs

   • Create a Conda environment for Nextflow:

   conda create --prefix /home/$USER/a/conda_envs/nextflow python=3 bioconda::nextflow  -y

5. Verify the Nextflow installation:

   conda activate /home/$USER/a/conda_envs/nextflow
   nextflow -h
   nextflow -v

If no errors were encountered, Nextflow is installed correctly.