clarkenj/abide-fmriprep-slurm

Create SLURM scripts for preprocessing ABIDE 1 and 2 wih fMRIPrep

abide-fmriprep-slurm

Create SLURM scripts for preprocessing ABIDE 1 and 2 wih fMRIPrep. The process was done on Compute Canada Beluga.

Dependency

• datalad
• fmriprep-slurm

Retrieving data

A BIDS competable version of the ABIDE dataset are avalible through datalad repository

Generating slurm jobs with fMRIPrep slurm

1. Set up your environment following this tutorial

2. Run ./generate_slurm_script.sh. It will call generate_slurm_script.sh; which is a modified version of singularity_run.bash from fmriprep-slurm.

   You need: root of output directory, path to bids dataset, output derivative directory name.

   A minimal use case of this version:

   bash ./fmriprep_slurm_singularity_run.bash \
           ${OUTPUT_ROOT_PATH} \
           ${BIDS_DATASET_PATH} \
           derivatives
   

3. To submit jobs for each dataset, the command to use will be similar to this.

   find "$OUTPUT_DIR"/.slurm/smriprep_sub-*.sh -type f | while read file; do sbatch "$file"; done
   

4. Archive the dataset. See this page

Timeseries extraction

TBA