De novo design of protein target specific scaffold-based Inhibitors via Reinforcement Learning

Mridula Bontha, Andrew McNaughton, Carter Knutson, Jenna Pope, Neeraj Kumar

Keywords: Protein target specific molecular design, reinforcement learning, Graph Neural Networks, Lead molecule optimization, Drug Discovery, and Protein-ligand interaction.

Abstract: Efficient design and discovery of target-driven molecules is a critical step in facilitating lead optimization in drug discovery. Current approaches to develop molecules for a given protein target are intuition-driven, hampered by slow iterative design-test cycles due to computational challenges in utilizing 3D structural data, and ultimately limited by the expertise of the chemist – leading to bottlenecks in molecular design. In this contribution, we propose a novel framework, called 3D-MolGNN_RL, coupling reinforcement learning (RL) to a deep generative model based on 3D-Scaffold to generate target candidates specific to a protein pocket building up atom by atom from the starting core scaffold. 3D-MolGNN_RL provides an efficient way to optimize key features by multi-objective reward function within a protein pocket using parallel graph neural network models. The agent learns to build molecules in 3D space while optimizing the binding affinity, potency, and synthetic accessibility of the candidates generated for the SARS-CoV-2 Main Protease.

One-sentence Summary: We propose a new method, known as 3D-MolGNN_RL, that not only incorporates the protein structure into the RL loop, but also considers the 3D structures of the generated compound build by placing atom by atom in the 3D space.