Skip to content


Adapted from

Code related to: O. Mendez-Lucio, M. Ahmad, E.A. del Rio-Chanona, J.K. Wegner, A Geometric Deep Learning Approach to Predict Binding Conformations of Bioactive Molecules, Nature Machine Intelligence volume 3, pages1033–1039 (2021)

Open access preprint available here

Getting Started

Main Requirements:

  • PyTorch = 1.10.0
  • CUDA Toolkit = 11.3
  • Python = 3.6.9
  • RDKIT = 2019.09.1


  • create Conda environment
conda create -n "current" python=3.6.9
  • activate newly created Conda environment
conda activate current
conda install pytorch==1.10.0 torchvision==0.11.0 torchaudio==0.10.0 cudatoolkit=11.3 -c pytorch -c conda-forge
  • if not yet available in your system, load CUDA 11.3
ml load CUDA/11.3.1
  • install PyTorch scatter, sparse and geometric
pip install torch-scatter -f
pip install torch-sparse -f
pip install torch-geometric
  • uninstall PyTorch spline
pip uninstall torch-spline-conv
  • uninstall PyMesh (necessary to generate .ply files)
wget --no-check-certificate
pip install pymesh2-0.2.0-cp36-cp36m-linux_x86_64.whl
git clone
cd ChemBench
pip install -e .
  • install Trimesh
conda install -c conda-forge trimesh
  • if not yet available in your system, fixed issues with the version of libstdc
module load GCC/9.3.0
  • install other dependencies
conda install Biopython
conda install cmake
conda install automake
conda install bison
conda install flex
conda install -c anaconda swig
conda install -c conda-forge apbs
conda install -c conda-forge pdb2pqry
  • install AmberTools to replace reduce3.34 (for protonation) - since it is deprecated and no longer available, it is now included in AmberTools
conda install -c conda-forge ambertools
  • install requirements.txt
pip install -r requirements.txt
  • install RDKIT
conda install -c conda-forge rdkit=2019.09.1


  1. Clone the repo
git clone
  1. Move into the project folder and update submodules
cd DeepDock
git submodule update --init --recursive
  1. Install prerequisite packages
pip install -r requirements.txt
  1. Install DeepDock pacakge
pip install -e .


  • navigate to the location of your apbs-pdb2pqr/pdb2pqr installation and run the Python (2.7) script to link with your current Conda environment. For more information, refer to the Dockerfile (lines 60 to 72)

sh [your conda environment WORK DIRECTORY]/install/apbs-pdb2pqr/pdb2pqr python2.7 scons/ install PREFIX="[your conda ENVIRONMENT PATH]/bin/pdb2pqr"

  • move the "multivalue" file to your Conda envirnoment path
 cp multivalue [your conda environment path]/share/apbs/tools/mesh/multivalue
  • setup necessary environment variables with the tools and respective paths

sh export MSMS_BIN=[your conda environment path]/bin/msms export APBS_BIN=[your conda environment path]/bin/apbs export PDB2PQR_BIN=[your conda environment path]/bin/pdb2pqr/ export MULTIVALUE_BIN=[your conda environment path]/share/apbs/tools/mesh/multivalue


You can get training and testing data following the next steps.

  1. Move into the project data folder
 cd DeepDock/data
  1. Use the following line to download the preprocessed data used to train and test the model. This will download two files, one containing PDBbind (2.3 GB) used for training and another containing CASF-2016 (32 MB) used for testing. These two files are enough to run all examples.
  1. In case you want to reproduce all results of the paper you will need to download the complete CASF-2016 set (~1.5 GB). You can do so with this command line from the data folder.

Example Usage

Usage examples can be seen directly in the Jupyter Notebooks included in the repo. We added examples for: