TIGR-PURR Quickstart Tutorial

This tutorial is meant to run you through a simple example of running TIGR-PURR pipelines on your computer using data from our internal SPINS study.

Please refer to Getting started for more detailed information about running TIGR-PURR pipelines.

Setting up the Quickstart Tutorial

When we’re running a TIGR-PURR pipeline the first thing to do is to organize your project directory. We’ll do a simple folder in scratch with the following command:

mkdir /scratch/<YOU>/nextflow_quickstart
cd /scratch/<YOU>/nextflow_quickstart

Note

If you want to go through this tutorial on the SCC you must add:

/KIMEL/tigrlab/

Before every path!

Next we’ll make a text file containing a list of participants from the SPINS study. For this tutorial we’ll only run a subset of participants since running all participants will take too much time:

touch sublist

The sublist text file will contain a list of participants (one-per-line) that we want to run, enter the following into sublist:

sub-CMH0144
sub-MRP0136
sub-MRC0021

Finally, load in the Nextflow module:

module load nextflow/19.04.1

Note

If you’re on the SCC you must load in our Kimel modules first before loading in nextflow:

module load /KIMEL/quarantine/modules/quarantine

Note

This will create a directory in:

/scratch/<YOU>/nextflow_work/

Which you should clean. For more information see Cleaning the Nextflow Work folder

Dry-Runs of Pipelines

In this quickstart tutorial we’ll be running a Dry-run of MRIQC on a bunch of participants from the SPINS study located in the archive. A dry-run is a way of running pipelines without doing actual computation.

Note

See :ref:`dryrun’ for more details about using a dry-run

Running the pipeline

Once you have a BIDS-app, subject list, and invocation ready to go you can run a pipeline! Let’s run it on the local system (your own computer) first, but feel free to try other profiles!:

nextflow /archive/code/tigrlab_nextflow/bids.nf \
        -c /archive/code/tigrlab_nextflow/nextflow_config/mriqc-0.14.2.nf.config \
        --bids /archive/data/SPINS/data/bids \
        --out /scratch/<YOU>/nextflow_quickstart \
        --subjects /scratch/<YOU>/nextflow_quickstart/sublist \
        --invocation /archive/code/boutiques_jsons/invocations/dryrun_mriqc-0.14.2_invocation.json \
        -profile local

The process will run the dry-run version of MRIQC in parallel automatically!

Note

Feel free to try different profiles like “kimel” or “scc”. For more information on profiles see: Running Pipelines on Other Systems (SCC/Scinet/Local)

Cleaning up your TIGR-PURR run

Once you’re finished running a pipeline, you need to clean out the Nextflow working directory. By default the working directory is found by examining the $NXF_WORK environment variable:

echo $NXF_WORK

To clean it out you can simply type in:

clean_nxf

Which will clean out the subfolders in this directory.

Note

For more information on cleaning working directories check out Cleaning the Nextflow Work folder