Design and molecular mechanism investigation of ALK inhibitors based on virtual screening and structural descriptor modeling.

Abstract

To address the challenges of target specificity and drug resistance in Anaplastic lymphoma kinase (ALK) inhibition, this study conducted a virtual screening of the BindingDB database, yielding 711 potential ALK inhibitors. Four QSAR models were established using structural clustering and machine learning to elucidate structure-activity relationships. Through substituent fragment optimization, 72 highly active compounds were designed, among which four promising candidates were identified based on ADMET predictions, retrosynthetic analyses and molecular docking analyses. Molecular dynamics simulations and binding free energy calculations further characterized their binding mechanisms. These findings provide a theoretical framework for the rational design of next-generation ALK inhibitors.