Quantifying Dynamic Hotspots in Protein-Ligand Interactions: A Computational Framework for Enhancing Drug Discovery

Protein-ligand interactions (PLIs) are central to drug discovery, yet traditional computational methods often overlook the dynamic nature of binding sites. This thesis presents a comprehensive computational framework that integrates molecular dynamics (MD) simulations with statistical analyses to identify and quantify dynamic binding hotspots, including cryptic and allosteric pockets. Utilizing a curated dataset of 100 high-resolution protein-ligand complexes, key dynamic descriptors—root mean square deviation (RMSD), solvent-accessible surface area (SASA), and hydrogen bond occupancy—were systematically analyzed to assess binding stability and pocket druggability. Results demonstrate that stable ligand binding is characterized by low RMSD (<2 Å), consistent SASA values, and high hydrogen bond occupancy (>70% duration). Furthermore, the enrichment of charged residues (Asp, His, Arg) in binding pockets underscores the role of electrostatics in interaction specificity. The study proposes reproducible quantitative benchmarks for validating PLIs, addressing critical gaps in pose validation and target prediction. This framework advances the accuracy and reliability of early-stage drug discovery, enabling more precise identification of therapeutically relevant binding sites.