READII-2-ROQC RADCURE for ORCESTRA

Running pipeline

Dependencies

Creating conda environment with snakemake

conda create -n snakemake -c conda-forge -c bioconda snakemake # (optional: singularity)

Running using GCP (Google Cloud Platform)

The bucket used is (GCP Project ID = orcestra-388613):

snakemake --profile workflow/profiles/gcp --kubernetes

Using Singularity

• containerization technology that allows users to run lightweight, isolated environments for specific use-cases.
• initially developed for Linux
• using on MacOS requires a linux VM (https://docs.sylabs.io/guides/3.0/user-guide/installation.html#mac)

snakemake -c1 --use-singularity

Using conda environments

snakemake -c1 --use-conda

Creating directed acyclic graph (DAG)

snakemake -c1 --dag | dot -Tsvg > dag.svg

Testing individual scripts

image_dir="orcestradata/radiomics/radcure_test_sample/images/RADCURE-0020"
segmentation_dir="orcestradata/radiomics/radcure_test_sample/images"
output_dir="results/radiomic_output"
python scripts/radiomic_extraction/radiogenomic_pipeline.py \
    --name "snakemake_RADCURE_pipeline" \
    --experiment "snakemake_RADCURE" \
    --image_dir ${image_dir} \
    --segmentation_dir ${segmentation_dir} \
    --output_dir ${output_dir} \
    --segmentation_modality RTSTRUCT \
    --roi_names GTVp.* \
    --pyrad_param_file scripts/radiomic_extraction/pyradiomics/pyrad_settings/settings_original_allFeatures.yaml \
    --quality_checks True