Virtual screening of human pancreatic lipase inhibitors using an integrated transfer learning and molecular dynamics simulation.

Human pancreatic lipase (hPL), a major enzyme involved in dietary lipid digestion is a crucial therapeutic target for the management of obesity. Therefore, it is of interest to identify hPL inhibitors using an integrated transfer learning and molecular dynamics simulation. Transfer learning predicted 17,309 active candidates based on the pIC50, from which prenylated flavanonol was identified as the top-ranked compound with a binding affinity of 9.85 kcal/mol in comparison to reference inhibitor orlistat using Glide XP docking. Moreover, 200ns molecular dynamics simulations confirmed the stability of the protein-ligand complex through sustained interactions with key catalytic residues, including Ser152 and Asp79. Thus, data shows that prenylated flavanonol as a promising natural inhibitor of hPL.