RVGrow A Fragment Efficiency-Based Scoring Strategy for Rational and Efficient Fragment-to-Lead Optimization

Fragment-based drug design (FBDD) seeks to optimize fragment hits into potent lead compounds by maximizing their three-dimensional interactions with the target binding site. However, designing efficient fragment-to-lead growth strategies remains a core challenge in de novo drug discovery. Here, we introduce RVGrow, a novel algorithm incorporating a Fragment Efficiency Score (FE-score) that guides rational fragment elaboration based on interaction efficiency. RVGrow was systematically applied to evaluate the correlation between FE-score and experimental inhibition data across multiple protein targets. Furthermore, its integration with deep generative models demonstrates RVGrow's adaptability and utility in AI-driven molecular design pipelines. The work presents RVGrow in the context of fragment-to-lead optimization, highlighting its use of interpretable, efficiency-based scoring.