FREED++: Improving RL Agents for Fragment-Based Molecule Generation by Thorough Reproduction

TL;DR

This paper reproduces and improves the FREED RL model for molecule generation, fixing bugs, simplifying the approach, and demonstrating superior docking scores over current methods in protein-targeted drug design.

Contribution

It provides a thorough reproduction, bug fixes, simplification, and extensive evaluation of the FREED model, enhancing its performance and reliability.

Findings

Fixed multiple implementation bugs in FREED.

Simplified the model while improving quality.

Achieved superior docking scores compared to state-of-the-art methods.

Abstract

A rational design of new therapeutic drugs aims to find a molecular structure with desired biological functionality, e.g., an ability to activate or suppress a specific protein via binding to it. Molecular docking is a common technique for evaluating protein-molecule interactions. Recently, Reinforcement Learning (RL) has emerged as a promising approach to generating molecules with the docking score (DS) as a reward. In this work, we reproduce, scrutinize and improve the recent RL model for molecule generation called FREED (arXiv:2110.01219). Extensive evaluation of the proposed method reveals several limitations and challenges despite the outstanding results reported for three target proteins. Our contributions include fixing numerous implementation bugs and simplifying the model while increasing its quality, significantly extending experiments, and conducting an accurate comparison with current state-of-the-art methods for protein-conditioned molecule generation. We show that the resulting fixed model is capable of producing molecules with superior docking scores compared to alternative approaches.