Abstract
Drug discovery faces increasing challenges in identifying novel drug candidates satisfying multiple stringent objectives, such as binding affinity, protein target selectivity, and drug-likeness. Existing optimization methods struggle with the complexity of handling numerous objectives, limiting advancements in molecular design as most algorithms are only effective for up to four optimization objectives. To overcome these limitations, the study introduces the Pareto Monte Carlo Tree Search Molecular Generation (PMMG) method, leveraging Monte Carlo Tree Search (MCTS) to efficiently uncover the Pareto Front for molecular design tasks in high-dimensional objective space. By utilizing simplified molecular input line entry system (SMILES) to represent molecules, PMMG efficiently navigates the vast chemical space to discover molecules that exhibit multiple desirable attributes simultaneously. Numerical experiments demonstrate PMMG's superior performance, achieving a remarkable success rate of 51.65% in simultaneously optimizing seven objectives, outperforming current state-of-the-art algorithms by 2.5 times. An illustrative study targeting Epidermal Growth Factor Receptor (EGFR) and Human Epidermal Growth Factor Receptor 2 (HER2) highlights PMMG's ability to generate molecules with high docking scores for target proteins and favorable predicted drug-like properties. The results suggest that PMMG has the potential to significantly accelerate real-world drug discovery projects involving numerous optimization objectives.