First, create a conda/mamba environment:
conda create --name pipeline python=3.10 pip
Install requirements:
conda install -c conda-forge openmm openff-toolkit openff-interchange openff-units openff-utilities openff-forcefields lomap2
conda install -c openbiosim sire==2024.3.0
and also install the requirements listed:
pip install -r requirements.txt
Clone the correct BioSimSpace version (https://github.com/michellab/BioSimSpace/releases/tag/RBFE-benchmark) and install this in the environment.
Follow the instructions there to install the branch.
To use flex align, kcombu also needs to be installed:
conda install openbiosim::kcombu_bss
Finally, go to the python directory and install the pipeline in the environment using:
python setup.py install
To run additional network analysis using either FreeEnergyNetworkAnalysis (https://github.com/michellab/FreeEnergyNetworkAnalysis) or MBARNet (https://gitlab.com/RutgersLBSR/fe-toolkit), please follow their instructions for installation and use. Default network analysis proceeds using cinnabar.
pipeline_notebooks - notebooks and scripts for starting and analysing the pipeline
python - contains all the code
other_scripts - contains additional notebooks.
After the initial setup of the pipeline (pipeline_notebooks folder), a main folder with the settings in execution_model should have been created. This contains the ligands.dat, network.dat, protocol.dat, and analysis_protocol.dat that can be edited as required. In the main folder, running: bash run_all_slurm.sh will start the entire pipeline for a slurm workload manager in a series of dependencies. The run_*_slurm scripts generated in the scripts folder may need to be adjusted depending on the slurm cluster and resources available. After all the runs are finished, they can be analysed in a jupyter notebook following the example in the pipeline_notebooks folder.
The workflow and scripts are outlined in the figure below:
