Start an interactive job, with a walltime of 2 hours, 2000MB of memory.
srun --pty -p interactive -t 0-02:0:0 --mem 2000MB -n 1 /bin/bash |
Create a working directory on scratch and change into the newly-created directory. For example, for user abc123, the working directory will be
mkdir /n/scratch/users/a/abc123/bwa cd /n/scratch/users/a/abc123/bwa |
Copy some test data following this page: Build Folder Structures From Sample Sheet for rcbio NGS Workflows page.
Copy the example bwa bash script:
module load gcc/6.2.0 python/2.7.12 rcbio/1.3.3 cp /n/app/rcbio/1.1/bin/bwa.sh . |
Now you can modify the options as needed. Please reference the bwa user manual if you have any questions.
To edit the bwa script:
nano bwa.sh |
To test the pipeline run the following command. Jobs will not be submitted to the scheduler.
runAsPipeline "bwa.sh -r hg38" "sbatch -p short --mem 6G -t 2:0:0 -n 1" useTmp # this is a test run |
To run the pipeline:
runAsPipeline "bwa.sh -r hg38" "sbatch -p short --mem 6G -t 2:0:0 -n 1" useTmp run 2>&1 | tee output.log # notice here 'run 2>&1 | tee output.log' is added to the command |
To understand how 'runAsPipeline' works, how to check output, how to re-run the pipeline, please visit: Run Bash Script As Slurm Pipeline
To instead run the workflow on your own data, transfer the sample sheet to your local machine following this wiki page and modify the sample sheet. Then you can transfer it back to O2 under your account, then go to the build folder structure step.