High predictive QSAR models for predicting the SARS coronavirus main protease inhibition activity of ketone-based covalent inhibitors

ABSTRACT: In this research, a dataset including 29 ketone-based covalent inhibitors with SARS-CoV-1 3CLpro inhibition activity was used to develop high predictive QSAR models. Twenty-two molecules were put in train set and seven molecules in test set. By using stepwise MLR method for molecules in train set, four molecular descriptors including Mor26p, Hy, GATS7p and Mor04v were selected to build QSAR models. MLR and ANN methods were used to create QSAR models for predicting the activity of molecules in both train and test sets. Both QSAR models were validated by calculating several statistical parameters. R2 values for the test set of MLR and ANN models were 0.93 and 0.95, respectively, and RMSE values for their test sets were 0.24 and 0.17, respectively. Other calculated statistical parameters (esp

SUBMITTER: Sepehri B 

PROVIDER: S-EPMC8547569 | biostudies-literature | 2021 Oct

REPOSITORIES: biostudies-literature