Advanced AI‐Driven in Silico Approaches for the Rapid Discovery of Potent BCL‐2 Inhibitors in Cancer Therapy

Abstract Cancer remains a leading cause of mortality worldwide, necessitating novel therapeutic approaches. This study utilized advanced in silico methods and artificial intelligence to identify new BCL‐2 inhibitors, a key apoptosis‐regulating protein often overexpressed in cancer cells. Machine learning‐based screening of 42 million Mcule compounds identified promising candidates, evaluated through molecular docking and dynamics simulations for BCL‐2 binding scores and stability. MetaCore/MetaDrug assessed drug‐likeness, QSAR predicted anticancer potential. Three promising molecules—Molecule #30, #73, and #75—were identified, consistently demonstrating superior or comparable binding scores to Venetoclax. For 1, 10, and 100 ns MD simulations, Molecule #30 scored −120.30, −122.02, and −122.68 kcal/mol; Molecule #73 scored −124.78, −125.62, and −125.42 kcal/mol; and Molecule #75 scored −120.02, −122.14, and −124.04 kcal/mol, respectively. Venetoclax, the reference drug, scored −100.66, −106.02, and −124.95 kcal/mol. These findings underscore the preclinical promise of these candidates, offering new opportunities for targeted cancer therapies and advancing oncology research.