Search

Type-2 diabetes mellitus (T2DM), is considered being a multifaceted disorder that presents a major therapeutic problem. Despite the widespread efforts in managing T2DM, drug discovery therapies remain insufficient. There are several new and vital antidiabetic drug agents under examination of many research clusters. Computational tools have influenced drug discovery at several levels. Hence, this study aimed to explain the molecular interactions via an in silico simulation approach between Aframomum melegueta seed-derived compounds as antidiabetic agents against carbohydrate metabolism regulatory target proteins (DPP-4, GSK-3, and GLP-1) involved in T2DM. Pharmacophore modeling, Auto-QSAR, induced-fit docking (IFD) simulation, free energy of binding affinity through the MM-GBSA method, and molecular docking were used to sort out the compounds as well as determining their drug ability. The results revealed the pharmacophore models created from the structures of DPP-4, GSK-3, and GLP-1 could determine the lead candidates with a restorative ability rate found near 100% of the actives in the comprehensive decoy database ranking. We predicted the pIC50 values of the compounds via a learning-based machine that was created by auto-QSAR. We authenticated the created model to confirm its analytical procedure. The greatest models attained for DPP-4, GSK-3, and GLP-1 were kpls_desc_35 (R2 = 0.73 and Q2 = 0.73), kpls_desc_4 (R2 = 0.72 and Q2 = 0.66), and kpls_radical_14 (R2 = 0.88 and Q2 = 0.64) and these confirmed models were used to predict the bioactivities of the hit candidates. While several isolated bioactive compounds from Aframomum melegueta revealed a better binding affinity score, binding free energy along with observance of RO5, only three compounds (ethinyl estradiol, 8-gingerol, and zingerone) revealed a more satisfactory pIC50. Thus, these compounds can be novel candidates for drug discovery in diabetes treatment.