Development of QSAR model using machine learning and molecular docking study of polyphenol derivatives against obesity as pancreatic lipase inhibitor

In developed countries and developing countries, obesity/overweight is considered a major problem, in fact, it is now recognized as a major metabolic disorder. Additionally, obesity is connected with other metabolic diseases, including cardiovascular disorders, type 2 diabetes, some types of cancer, etc. Therefore, the development of novel drugs/medications for obesity is essential. The best target for treating obesity is Pancreatic Lipase (PL), it breaks 50-70% triglycerides into monoglycerol and free fatty acids.The major aim of this in silico study is to generate a QSAR model by using Multiple Linear Regression (MLR) and to inhibit pancreatic lipase by polyphenol derivatives mainly flavonoids, plant secondary metabolites shows good inhibitory activity against PL, maybe with less unpleasant side effects.In this in silico study, a potent inhibitor was found through calculating drug likness, QSAR (Quantitative structure-activity relationship) and molecular docking. The docking was performed in Maestro 12.0 and the ADME (absorption, distribution, metabolism, and excretion) properties (drug-likeness) of compounds/ligands were predicted by the Qikprop module of Maestro 12.0. The QSAR model was developed to show the relationship between the chemical/structural properties and the compound's biological activity. We have found the best interaction between pancreatic lipase and flavonoids. The best docked compound is Epigallocatechin 3,5,-di-O-gallate with docking score -10.935 kcal/mol .All compounds also show drug-likeness activity.The developed model has satisfied all internal and external validation criteria and has square correlation coefficient (r2) 0.8649, which shows its predictive ability and has good acceptability, predictive ability, and statistical robustness.Communicated by Ramaswamy H. Sarma.