부산대학교 도서관

Severe Acute Respiratory Syndrome Coronavirus2 (SARSCOV2) is the virus that causes a respiratory disease called coronavirus (COVID-19). The symptoms of SARSCOV2 are fatigue, shortness of breath and cough that lasts more than four weeks. This paper aims for discovering active compounds for a particular target thatisSARSCOV2. Anattempttodiscover a novel compound based on the quantitative structure-activity relationship (QSAR) and drug-likeness evaluation was taken in the experiment. However, conventional laboratory processes are costlier and more time-consuming. Thus, it is extremely appropriate to develop an alternative method for predicting to inhibition of SARSCOV2. A non-redundant data set of 98 large molecules of SARSCOV2 inhibitory activity is used. Random forest (RF), sup- port vector regressor (SVR), and linear regression (LR) are used to predict the -log (IC50) value. The predictive performances for RF, SVR and LR in terms of R- squared, mean absolute error (MAE),and root mean square error(RMSE)are(0.77,0.89,0.80), (0.50, 0.26, 0.40) and (0.34, 0.17, 0.23), respectively. It can be concluded that the data preprocessing and the proposed QSAR model seems to be high-throughput tool for detecting novel inhibitors against SARSCOV2.