CanSig-benchmark

Complete documentation and tutorials: https://boevalab.github.io/CanSig-benchmark/

Running the benchmark

Installing the enviorments

This repository provides conda environments for reproducing our benchmarking. We provide two installation options for different reproducibility needs.

Option 1: Exact Reproducibility

Use environments from envs/with_build/:

conda env create -f envs/with_build/CanSig-R.yml
conda env create -f envs/with_build/CanSig-python.yml
conda env create -f envs/with_build/cansig-benchmark.yml

Option 2: Flexible Installation

Use environments from envs/without_build/:

conda env create -f envs/without_build/CanSig-R.yml
conda env create -f envs/without_build/CanSig-python.yml
conda env create -f envs/without_build/cansig-benchmark.yml

The whole benchmark is run with the cansig-benchmark environment and snakemake activates the other environments.

conda activate cansig-benchmark

Downloading the data

To run the benchmark, first download the data from the 3ca.

python ccafetcher.py

Install the necessary environments. Then use snakemake to run the benchmark.

snakemake --configfile config.yml -c <n-threads> --use-conda

Changes made to the 3CA.

• Fix the metadata of Neftel et al. by changing the technology to 10x for some samples.
• Converted the .rds in Couturier et al. to .mtx.
• The dataset from Couturier et al. was subsetted to cells that where identified as IDH WT in genetic_hormonal_features and in histology as GBM.
• The dataset from Yuan et al. was subsetted to GBM patients.