SampleDock

• Molecular design framework that merges generative AI and molecular docking (manuscript and webpage)

Installation:

Option 1

Installation using Anaconda

Note: The sampledock conda package is only distributed for linux systems. You will need to use Option 2 for installation on Mac OS and Windows

conda create -n sampledock python=3.7
conda activate sampledock
conda install sampledock -c AaronFrankLab -c conda-forge -c bioconda -c pytorch -c tmap

Option 2

If you wish to install from source, the general requirement is listed below.

Main dependencies:

  - python=3.7
  - pytorch
  - cudatoolkit (optional only for cuda enabled device)
  - scipy
  - rdkit
  - rxdock (or rdock)

Data visualizations tools:

  - tmap
  - faerun
  - mhfp

You can clone this repo and install the required python packages with conda environment.yml

conda env create -f environment.yml
conda activate sampledock
python setup.py install

If you are using Anaconda and install the required packages with the environment.yml, the docking program, rDock, will be installed as the package rxdock. This is precompiled executables released by RxDock. rxdock has some commandline argument changes from rDock. The original rDock can be installed following the instruction: http://rdock.sourceforge.net/installation/.

Usage:

To run Sample and Dock, first specify the hyperparameters in hyper.param.

Then, python -m sampledock hyper.param

JTVAE Model:

The JTVAE model is developed by Jin, W., Jaakkola, T. et al. and retireved from git: https://github.com/wengong-jin/icml18-jtnn. The default JTVAE generation model, moses-h450z56, is supplied with and trained with MOSES dataset.

The python scripts for JTVAE are modified for compatiblity with python 3.7.

Target Specific Libraries:

SARS-CoV-2: https://github.com/atfrank/SARS-CoV-2