Diabetes mellitus is a serious health disease that affects people all over the world. The number of persons identified with diabetes mellitus rises each year. α -Glucosidase is a digestive enzyme used to control diabetes mellitus. The searching for new potent α-glucosidase inhibitors capable of delaying carbohydrate digestion in the human body is an important strategy towards control of diabetes mellitus. In this work, a series of quinoline-based Schiff base derivatives already identified as α-glucosidase inhibitory activity was studied by using 2D/3D-QSAR approach. The best HQSAR/A-B-C-H-Ch-DA and CoMSIA/SEDA models were constructed using thirteen molecules in the training set, resulting in favorable values of Q2 (0.834 and 0.607), and high values of R2 (0.985 and 0.912), respectively. The generated HQSAR/A-B-C-H-Ch-DA and CoMSIA/SEDA contour plots were precious for designing and enhancing the α-glucosidase inhibitory activity of quinoline-based Schiff base molecules. Considering these results, two novel α-glucosidase compounds were designed to possess significant activity. The newly suggested molecules showed good outcomes in the preliminary in silico ADME/Tox evaluations. Molecular docking results revealed that the new designed inhibitors have a good stability in the active pocket of the studied receptor compared to voglibose, clinically used as an α-glucosidase inhibitor. MD simulation and MM-GBSA results confirmed the molecular docking outcomes. Finally, DFT analysis was useful in determining the most electrophilic and nucleophilic centers of the two designed α-glucosidase inhibitors.

Protein IDO1 (indoleamine 2,3-dioxygenase) occupies a critical position in the regulation of the immune system and is involved in cancer progression and the development of immune diseases. Being a therapeutic target for such critical diseases, we aimed to investigate the IDO1 inhibition activity of thirty-nine triazole derivatives using a quantitative structure-activity relationship. The dataset was under principal component analysis, multiple linear regression, and multiple non-linear regression from which two models were generated. The best 2D-QSAR model was generated using linear regression, demonstrating a determination coefficient of R2=0.680, a good acceptable internal cross-validated coefficient of R2cv=0.700, an error of MSE=0.074, and a good predictive potential of R2test=0.809. The QSAR model was further investigated using the applicability domain, which showed that all molecules were within the applicability domain, hence the absence of an outlier. Overall, the obtained results provide a reliable and highly predictive model for the design and prediction of new IDO1 inhibitors thereby influencing cancer progression and autoimmune disease development.