Hit and Lead Discovery with Explorative RL and Fragment-based Molecule Generation

TL;DR

This paper introduces FREED, a novel reinforcement learning framework that combines fragment-based molecule generation and error-prioritized experience replay to generate chemically realistic molecules with high docking scores for drug discovery.

Contribution

FREED is the first RL method to integrate fragment-based generation with error-prioritized experience replay for improved drug molecule design.

Findings

FREED outperforms existing methods in generating higher quality molecules.

The model achieves state-of-the-art docking scores on two of three targets.

Ablation studies confirm the effectiveness of predictive error-PER in improving performance.

Abstract

Recently, utilizing reinforcement learning (RL) to generate molecules with desired properties has been highlighted as a promising strategy for drug design. A molecular docking program - a physical simulation that estimates protein-small molecule binding affinity - can be an ideal reward scoring function for RL, as it is a straightforward proxy of the therapeutic potential. Still, two imminent challenges exist for this task. First, the models often fail to generate chemically realistic and pharmacochemically acceptable molecules. Second, the docking score optimization is a difficult exploration problem that involves many local optima and less smooth surfaces with respect to molecular structure. To tackle these challenges, we propose a novel RL framework that generates pharmacochemically acceptable molecules with large docking scores. Our method - Fragment-based generative RL with Explorative Experience replay for Drug design (FREED) - constrains the generated molecules to a realistic and qualified chemical space and effectively explores the space to find drugs by coupling our fragment-based generation method and a novel error-prioritized experience replay (PER). We also show that our model performs well on both de novo and scaffold-based schemes. Our model produces molecules of higher quality compared to existing methods while achieving state-of-the-art performance on two of three targets in terms of the docking scores of the generated molecules. We further show with ablation studies that our method, predictive error-PER (FREED(PE)), significantly improves the model performance.