ReACT-Drug: Reaction-Template Guided Reinforcement Learning for de novo Drug Design

TL;DR

This work introduces ReACT-Drug, a fully integrated, target-agnostic molecular design framework based on Reinforcement Learning, which highlights the potential of integrating structural biology, deep representation learning, and chemical synthesis rules to automate and accelerate rational drug design.

Abstract

De novo drug design is a crucial component of modern drug development, yet navigating the vast chemical space to find synthetically accessible, high-affinity candidates remains a significant challenge. Reinforcement Learning (RL) enhances this process by enabling multi-objective optimization and exploration of novel chemical space - capabilities that traditional supervised learning methods lack. In this work, we introduce \textbf{ReACT-Drug}, a fully integrated, target-agnostic molecular design framework based on Reinforcement Learning. Unlike models requiring target-specific fine-tuning, ReACT-Drug utilizes a generalist approach by leveraging ESM-2 protein embeddings to identify similar proteins for a given target from a knowledge base such as Protein Data Base (PDB). Thereafter, the known drug ligands corresponding to such proteins are decomposed to initialize a fragment-based search space, biasing the agent towards biologically relevant subspaces. For each such fragment, the pipeline employs a Proximal Policy Optimization (PPO) agent guiding a ChemBERTa-encoded molecule through a dynamic action space of chemically valid, reaction-template-based transformations. This results in the generation of \textit{de novo} drug candidates with competitive binding affinities and high synthetic accessibility, while ensuring 100\% chemical validity and novelty as per MOSES benchmarking. This architecture highlights the potential of integrating structural biology, deep representation learning, and chemical synthesis rules to automate and accelerate rational drug design. The dataset and code are available at https://github.com/YadunandanRaman/ReACT-Drug/.

Figures (6)

• Figure 1: De novo molecular drug discovery pipeline workflow. The system initializes with a target protein, identifies similar proteins using ESM-2 embeddings, fragments their ligands, and uses a PPO agent to iteratively optimize molecules over 15-step episodes. The agent selects chemical transformations from a drug-relevant template library, with rewards based on binding affinity (AutoDock Vina), drug-likeness (QED), synthetic accessibility, and novelty.
• Figure 2: Illustrative example of the ReACT-Drug pipeline.
• Figure 3: Physicochemical property profiles of generated molecules across six diverse targets. While Molecular Weight (MW) remains consistent across targets (clustered at $\approx 0.8$ normalized units), other properties like Hydrogen Bond Acceptors (HBA) and Donors (HBD) vary significantly (for example, high HBA demand for 5-HT1B vs. high HBD for DRD2), demonstrating the model's capacity to adapt to target-specific constraints. Values are min-max normalized for visual comparison.
• Figure 4: Comparison of (a) QED scores and (b) Synthetic Accessibility scores across six protein targets. Higher QED indicates better drug-likeness, while lower SA scores indicate easier synthesis.
• Figure 5: Binding affinity scores (kcal/mol) achieved by our target-agnostic framework across six diverse protein targets. More negative values indicate stronger predicted binding.