Installation

In order to run the ChromBPNet pipeline within this repository, you will need to first install the conda environment provided. All models must be run using NVIDIA A100 GPU nodes, so it is necessary to have the required NVIDIA drivers and CUDA already installed. Running nvidia-smi will tell you whether your environment is properly set up to use GPUs.

Local installation of the ChromBPNet environment:

# Clone repository 
git clone <repo-url> Pituitary_Chromatin_Atlas
cd Pituitary_Chromatin_Atlas

# Create environment
conda env create -f ChromBPNet_environment.yml

# Activate environment
conda activate ChromBPNet

For running the BPNet pipeline, example workflow files are provided in `BPNet_pipeline', but these are written from the bpnet-lite repository, with no changes made, so it is recommended to refer to their tutorial in order to execute any BPNet model training. Nevertheless, an environment is provided for quick installation and function use:

# Create environment
conda env create -f BPNet_environment.yml

# Activate environment
conda activate BPNet-lite