Running FASTQC with Slurm

Slurm is a resource manager. You can run your codes as Slurm scripts.

Sample Slurm script: fastqc_slurm_submit.sh

#!/bin/bash

#SBATCH -A hpc_training                    # account name (--account)
#SBATCH -p standard                        # partition/queue (--partition)
#SBATCH --nodes=1                          # number of nodes
#SBATCH --ntasks=1                         # 1 task – how many copies of code to run
#SBATCH --cpus-per-task=1                  # total cores per task – for multithreaded code
##SBATCH --mem=3200                        # total memory (Mb) *Note ##comment
#SBATCH -t 00:20:00                        # time limit: 20 min
#SBATCH -J fastqc-test                     # job name
#SBATCH -o fastqc-test-%A.out              # output file
#SBATCH -e fastqc-test-%A.err              # error file
#SBATCH --mail-user=dtriant@virginia.edu   # where to send email alerts
#SBATCH --mail-type=ALL                    # receive email when starts/stops/fails

__module purge__   # good practice to purge all modules
__module load__  fastqc/0.12.1

cd /project/rivanna-training/bioinformatics-hpc

mkdir fastqc-out-slurm

__fastqc__  -t 4 -o fastqc-out-slurm  ecoli-fastq/SRR258*fastq  # command for running software

Mail option output: The mail option notifies you when the job starts and completes.

Job started:

Slurm Job_id=4972539 Name=busco-test Began, Queued time 00:00:46

Job ended:

Operational Terminal Commands

To submit your Slurm script from the terminal, use the following command:

$ sbatch fastqc_slurm_submit.sh

Submitted batch job 4938712

For job monitoring, use the following command:

$ squeue -u user_id  # checks job status

PD - pending

R - running

CG - exiting

The Slurm script method should have the same output files as running the job interactively.