snake-cnv

Snakemake-based CNV calling pipeline using MrCaNaVar and mrsFAST

Installation

Get the code.

git clone --recursive https://github.com/huddlej/snake-cnv.git
cd snake-cnv

Install the latest Python 3 version of Miniconda for your operating system. After installing Miniconda, create a new conda environment with Snakemake installed.

conda create -n snake-cnv -c bioconda -c conda-forge python=3 conda snakemake pandas

If you want to download raw reads using Aspera, download and install the Aspera command line interface (CLI) or find the path to the ascp command on your cluster environment. Edit the aspera_key_path parameter in config.json to point to the absolute path of your Aspera key. This is usually something like <PATH TO ASPERA INSTALLATION>/etc/asperaweb_id_dsa.openssh.

Linux-specific installation

Build dependencies for the analysis pipeline.

make

Mac OS X installation

Install the OpenMP library with Homebrew and then build dependencies as follows.

brew install libomp
make OPENMP_CXX="g++ -lstdc++ -lz -lm -Xpreprocessor -fopenmp -lomp"

Running on a local machine

Run Snakemake on your local machine with conda environment enabled. The first time you run this will take longer than subsequent runs, since Snakemake will download and install all dependendies in a custom conda environment.

conda activate snake-cnv
snakemake -w 30 -j 2 --use-conda

Running on a SLURM cluster

conda activate snake-cnv
snakemake -w 30 -j 2 --use-conda --cluster-config cluster.json --cluster "sbatch --nodes=1 --ntasks=1 --mem={cluster.memory} --cpus-per-task={cluster.cores} --tmp={cluster.disk} --time={cluster.time} --job-name='{cluster.name}' --output='{cluster.stdout}' --error='{cluster.stderr}'"