Abstract
Salt-inducible kinases (SIK1, SIK2, and SIK3) regulate metabolism and immune responses, making them promising targets for inflammatory and autoimmune diseases. Understanding inhibitor selectivity among isoforms is crucial for therapeutic development. In this study, 44 compounds were investigated as SIK inhibitors using molecular modeling. A flexible treatment of the kinases via molecular dynamics (MD) simulations captured binding site conformational changes, followed by molecular docking to generate protein kinase (PK)-ligand complex models. Ligand orientations were validated against crystallographic data using LigRMSD and interaction fingerprints (IFPs). A genetic algorithm was applied to select conformations that maximize correlation between docking energies and biological activities, yielding R² values of 0.821, 0.646, and 0.620 for SIK1, SIK2, and SIK3, respectively. Our results highlight the importance of protein flexibility in achieving accurate correlations between docking energies and experimental pIC(50) values, enhancing inhibitor selectivity predictions.