dkoes / ReLeaSE

Deep Reinforcement Learning for de-novo Drug Design

ReLeaSE (Reinforcement Learning for Structural Evolution)

Deep Reinforcement Learning for de-novo Drug Design

Currently works only under Linux

This is a PyTorch implementation of Deep Reinforcement Learning for de-novo Drug Design aka ReLeaSE method.

REQUIREMENTS:

• Python 3.6 (we recommend Anaconda distribution)
• PyTorch 0.4 https://pytorch.org
• Modern NVIDIA GPU, compute capability 3.5 of newer.
• CUDA 9.0
• scikit-learn (http://scikit-learn.org/)
• rdkit (http://www.rdkit.org/)

Installation

Clone this repository into desired folder and install dependencies.

For use cases please refer to examples with several iPython notebooks

Citation

If you use this code or data, please cite:

ReLeaSE method paper:

Mariya Popova, Olexandr Isayev, Alexander Tropsha. Deep Reinforcement Learning for de-novo Drug Design. Science Advances, 2018, Vol. 4, no. 7, eaap7885. DOI: 10.1126/sciadv.aap7885