REINVENT with Quantum Chemistry Calculations

This repository includes:

• REINVENT and related packages in ./REINVENT and ./pkgs
• Initial ChEMBL prior network in ./data
• Calibration set along with TD-DFT results and references in ./calibration
• Results for generated molecules in ./results
• Scripts for setting up configurations and starting training for REINVENT in ./run

Installation

Note: The part of source codes related to job management and data parsing may need to be modified to fit your own system.

1. Clone this GitHub repo
2. Create a conda environment with reinvent.yml

$ conda env create -f reinvent.yml

Requirements

• SCScore: Synthetic complexity score prediction model
• AiZynthFinder: Retrosynthetic planning tool

Acknowledgments

• REINVENT 3.2: REINVENT model
• ReinventCommunity: Useful tutorials for REINVENT in jupyter notebooks