QSAR Study on Predicting DPP-IV Inhibitors as Anti-Diabetic Agent by using Genetic Algorithm-Support Vector Machine
Isman Kurniawan, Attariq Muhammad Kasfilla, Nurul Ikhsan · 2022
Diabetes mellitus is a major degenerative disease in the 21st century that causes nearly 95% of adults to be diagnosed with type II diabetes. One of the enzymes responsible for type II diabetes is Dipeptidyl peptidase-IV (DPP IV). Due to the limited number of inhibitors for type II diabetes, there is an urgent need to develop additional new DPP IV inhibitors. Along with technological developments, several research are involved in discovering and optimizing new DPP IV inhibitors, including implementing the quantitative structure-activity relationship (QSAR) method. Many studies have been performed to find a cure for diabetes mellitus using QSAR. The QSAR model have been commonly developed by using a wide range of machine learning algorithm, such as support vector machine (SVM). In this study, we aimed to develop a model to predict the activity of anti-diabetes. This research consists of two main steps, i.e., feature selection and prediction model development. The feature selection was carried out using the genetic algorithm (GA) method, and the prediction model was developed using SVM. Based on the results, we found that SVM with RBF kernel produces the best results with the value of accuracy and F-1 score are 0.9869 and 0.9871, respectively.