4. Setup projects with RUMP using HiPerGator

Login (5s)

Replace gatorlink_username in following command with your own username

ssh gatorlink_username@hpg2.rc.ufl.edu

Installing

See Installing

Direct to the RUMP folder (<1s)

replace /blue/djlemas/xinsongdu/jupyter_notebook/projects/RUMP with the location on your own machine

cd /blue/djlemas/xinsongdu/jupyter_notebook/project/RUMP/

Pull singularity image (5m)

mkdir -p work/singularity && singularity pull --name work/singularity/xinsongdu-lemaslab_rump-latest.img docker://xinsongdu/lemaslab_rump:latest

Copy the singularity image to the path pointed by singularity.cashDir (default value: "/blue/djlemas/share/containers/", which can be changed by modifying singularity.cashDir in nextflow.config)

cp work/singularity/xinsongdu-lemaslab_rump-latest.img /blue/djlemas/share/containers/

Load needed modules (10s)

ml nextflow singularity

Execute RUMP with provided sample data. (8 hours when using default configuration)

nextflow main.nf -profile singularity --input_dir_pos functional_test/sample_data/POS/ --input_dir_neg functional_test/sample_data/NEG --POS_design_path functional_test/sample_data/pos_design.csv --NEG_design_path functional_test/sample_data/neg_design.csv --container singularity

Execute RUMP to process your own data. (See executing time estimation)

1. Save your positive data files to data/POS/ and negative data to data/NEG/
2. Create design files for positve data and negative data, indicating the group of each file, save them to data/pos_design.csv and data/neg_design.csv. Sample design file can be found in data/sample_data/pos_design.csv and data/sample_data/neg_design.csv
3. Create a file named run.sh (or whatever you want but the extension should be .sh) including the following content. You can use a tool like vi for file creation.

#!/bin/bash
#SBATCH --job-name=nf_samples # You can change nf_samples to whatever name you want
#SBATCH --mail-type=ALL # Use ALL or NONE
#SBATCH --mail-user=xinsongdu@ufl.edu # Change this to your own email address
#SBATCH --ntasks=20 # Number of CPUs you allocate to your task
#SBATCH --mem=200gb # Number of memory you allocate to your task
#SBATCH --account=djlemas # UFRC account name
#SBATCH --qos=djlemas-b # Which quota of the account you want to use
#SBATCH --time=20:00:00 # Maximum running time of the task
#SBATCH --output=./logs/nf_samples%j.log # Log file of the task
pwd; hostname; date

# load modules
ml R
ml nextflow/19.10.0
ml singularity
SINGULARITYENV_TMPDIR=$(pwd)/tmp
export SINGULARITYENV_TMPDIR
TMPDIR=$(pwd)/tmp
export TMPDIR

nextflow main.nf -profile singularity
date

4. Process your data (replace run.sh with your own sbatch file name if the name is incorrect)

sbatch run.sh