knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  eval = FALSE
)

Create an environment with nextflow

cd /scratch/mblab/$USER

# launch an interactive session
srun --mem=20000 --cpus-per-task=1 -J interactive -p interactive --pty /bin/bash -l

# note: there are no promises that this package works. My suggestion is that you use your own installation of conda
ml miniconda

mkdir conda_envs

# this will create the environment directory in your scratch space (yes, it may be deleted. you'd have to make a new one when that happens), and installs nextflow into it
create create -p conda_envs/nextflow nextflow

Launch the environment and test nextflow

Not with conda activate nextflow. Because you created this outside of your $CONDA_HOME, you have to provide the path like so:

# do this if you are not already there
cd /scratch/mblab/$USER
srun --mem=20000 --cpus-per-task=1 -J interactive -p interactive --pty /bin/bash -l
ml miniconda


# activate the environment
source activate ./conda_envs/nextflow

nextflow run hello # note: might take a few minutes

Download the configuration files

You could put this in $HOME, if you want, to avoid having it garbage collected

git clone https://github.com/cmatKhan/configs.git

In the submission to nextflow, you will use one of the configurations in the subdirectory /conf of the repository above. Eventually this may be merged into the main fork of the configs repository, at which point you would just include the argument -p wustl_htcf in your nextflow command



cmatKhan/brentlabRnaSeqTools documentation built on Nov. 17, 2021, 5:47 a.m.