De novo Drug Design using Reinforcement Learning with Multiple GPT Agents

TL;DR

This paper introduces MolRL-MGPT, a reinforcement learning framework with multiple GPT agents that collaboratively generate diverse drug-like molecules, demonstrating promising results on benchmarks and SARS-CoV-2 inhibitor design.

Contribution

The paper presents a novel multi-agent reinforcement learning approach using GPT models for de novo drug design, enhancing molecular diversity and target specificity.

Findings

Effective in generating diverse molecules on GuacaMol benchmark

Successfully designed inhibitors against SARS-CoV-2

Demonstrates collaborative multi-agent reinforcement learning benefits

Abstract

De novo drug design is a pivotal issue in pharmacology and a new area of focus in AI for science research. A central challenge in this field is to generate molecules with specific properties while also producing a wide range of diverse candidates. Although advanced technologies such as transformer models and reinforcement learning have been applied in drug design, their potential has not been fully realized. Therefore, we propose MolRL-MGPT, a reinforcement learning algorithm with multiple GPT agents for drug molecular generation. To promote molecular diversity, we encourage the agents to collaborate in searching for desirable molecules in diverse directions. Our algorithm has shown promising results on the GuacaMol benchmark and exhibits efficacy in designing inhibitors against SARS-CoV-2 protein targets. The codes are available at: https://github.com/HXYfighter/MolRL-MGPT.