Getting Started

The CanSig-benchmark is a comprehensive benchmarking framework for cancer gene signature methods. This guide will walk you through setting up the environment and preparing the data needed to run the benchmarking pipeline.

Environment Setup

The base environment contains Snakemake, Apptainer, and other essential tools needed to orchestrate the benchmarking pipeline. All computational methods are containerized to ensure reproducibility across different systems.

Install the base environment that includes Snakemake and Apptainer:

conda env create -f envs/cansig.yml

Activate the base environment:

conda activate cansig

Data Preparation

After setting up the environment, you’ll need to download the cancer datasets from the Curated Cancer Cell Atlas (CCCA). These datasets form the foundation for the benchmarking analysis. For detailed instructions on downloading and processing the CCCA datasets, see the CCCA Data Download Guide.

Quick start for downloading a test dataset:

python ccafetcher.py --metadata 3ca_test.csv --download-dir data/downloads --dataset-dir data/raw --sample 0 --fetch

Running the Snakemake Pipeline

Once you have set up the environment and downloaded the necessary datasets, you can execute the benchmarking pipeline. The pipeline uses docker containers to ensure all methods run in isolated, reproducible environments.

Execute the pipeline using docker containers:

snakemake --configfile <path_to_your_config> --sdm apptainer -c <number_of_cores> -s <snakemake_file>

Replace <path_to_your_config> with your specific configuration file, <number_of_cores> with the number of CPU cores you want to allocate and <snakemake_file> with either signatures.smk or integration.smk depending on which part of the pipeline you wish to run.

Reproducing Paper Results

To reproduce the specific results from the CanSig-benchmark paper, use the provided configuration files. The pipeline includes two main components: signature analysis and integration analysis.

Signature Analysis Pipeline:

snakemake --configfile configs/signatures/<config>.yaml --sdm apptainer --apptainer-args "\\-\\-nvccli" -c <number_of_cores> -s signatures.smk

Integration Analysis Pipeline:

snakemake --configfile configs/integration/<config>.yaml --sdm apptainer --apptainer-args "\\-\\-nvccli" -c <number_of_cores> -s integration.smk

Replace <config> with the specific configuration file name you want to use. The --nvccli argument enables NVIDIA container support for GPU-accelerated methods.

Next Steps

1. Configure your analysis: Examine the configuration files in config/signatures/ and config/integration/ to understand the available options and tutorials.

2. Analyze results: Once done running your benchmark you can analyze the created data using the notebooks provided in figures/