Skip to content

Instantly share code, notes, and snippets.

@brantfaircloth
Last active August 29, 2015 14:16
Show Gist options
  • Save brantfaircloth/f70aba5c797eede81af8 to your computer and use it in GitHub Desktop.
Save brantfaircloth/f70aba5c797eede81af8 to your computer and use it in GitHub Desktop.
bcl2fastq demxuing on zcluster
  • create shell script to run configuration of job. Save the following as configure.sh:

    #!/bin/bash
    export LD_LIBRARY_PATH=/usr/local/boost/1.44.0/gcc447/lib:${LD_LIBRARY_PATH}
    
    cd /escratch4/tcg_tech2/tcg_tech2_Mar_03/test-ACN5P
    
    # create a basecalls directory to hold our output
    /usr/local/bcl2fastq/1.8.4/bin/configureBclToFastq.pl \
        --input-dir 150213_M02849_0082_000000000-ACN5P/Data/Intensities/BaseCalls/ \
        --out basecalls \
        --use-bases-mask Y*,I8,I8,Y* \
        --sample-sheet 150213_M02849_0082_000000000-ACN5P/Data/Intensities/BaseCalls/SampleSheet.csv
    
  • submit the job to the queuing software:

      qsub -q rcc-30d configure.sh
    
  • that should run and create the "basecalls" directory. look at the contents of the stderr file to make sure it's all green. If it's not, nuke the basecalls directory, fix the errors, and rerun the above:

    cat configure.sh.e7046793
    
  • looks good. now lets create the actual demuxing script in file named make.sh bcl2fastq doesn't run that much better with 8 threads than 4, so just use 4. You could change to 8 here (be sure to change in both places). Save the following in make.sh. You only need to invoke make once (for all files):

    #!/bin/bash
    export OMP_NUM_THREADS=4
    export LD_LIBRARY_PATH=/usr/local/boost/1.44.0/gcc447/lib:${LD_LIBRARY_PATH}
    
    cd /escratch4/tcg_tech2/tcg_tech2_Mar_03/test-ACN5P/basecalls
    make -j 4
    
  • submit the job to the queuing software:

      qsub -q rcc-30d -pe thread 4 make.sh
    
  • data are going to be in basecalls/Project_default/, where the "default" part of that name is the project name you used.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment