Predictive\nModeling and Drug Repurposing for Type-II\nDiabetes

Diabetes\nmellitus (DM) is a global health concern, and dipeptidyl\npeptidase-4 (DPP-4) is a key therapeutic target. The study used three\nmachine learning and deep learning models to predict potential DPP-4\ninhibitors using a curated data set of 6,750 compounds. The models\nincluded support vector machine (SVM), random forest (RF), naive Bayes\n(NB), and multitask deep neural network (MTDNN). The MTDNN model demonstrated\nstrong predictive performance, achieving 98.62% train accuracy and\n98.42% test accuracy for predicting DPP-4 inhibitors and a correlation\ncoefficient of 0.979 for training and 0.977 for the test data set,\nwith low training and test errors while predicting corresponding IC<sub>50</sub> values. The MTDNN model predicted potential inhibitors using\nan external data set of FDA-approved drugs, identifying 100 compounds.\nAmong these, five compounds stood out with promising molecular docking\nand dynamic profiles, suggesting their potential as repurposed drugs\nfor targeting DPP-4 and offering hope for the future of diabetes treatment.