Support vector regression-based QSAR models for prediction of antioxidant activity of phenolic compounds

Abstract

The Support vector regression (SVR) was used to investigate quantitative structure–activity relationships (QSAR) of 75 phenolic compounds with Trolox-equivalent antioxidant capacity (TEAC). Geometric structures were optimized at the EF level of the MOPAC software program. Using Pearson correlation coefficient analysis, four molecular descriptors [n(OH), Cosmo Area (CA), Core-Core Repulsion (CCR) and Final Heat of Formation (FHF)] were selected as independent variables. The QSAR model was developed from the training set consisting of 57 compounds and then used the leave-one-out cross-validation (LOOCV) correlation coefficient to evaluate the prediction ability of the QSAR model. Used Artificial neural network (ANN) and multiple linear regression (MLR) for comparing. The RMSE (root mean square error) values of LOOCV in SVR, ANN and MLR models were 0.44, 0.46 and 0.54. The RMSE values of prediction of external 18 compounds were 0.41, 0.39 and 0.54 for SVR, ANN and MLR models, respectively. The obtained result indicated that the SVR models exhibited excellent predicting performance and competent for predicting the TEAC of phenolic compounds.

Quantitative structural assessments of potential meprin β inhibitors by non-linear QSAR approaches and validation by binding mode of interaction analysis

“…The performance of the machine learning models and their learning capability were evaluated based on their squared correlation coefficient value and the corresponding root mean square error (RMSE) values and the Leave-one-out (LOO) crossvalidation process for the training set. 43,59 The predictive capability of these models for the training data was investigated by calculating their LOO-cross-validated R 2 (Q 2 ) values, whereas the external predictability was evaluated by the externally crossvalidated R 2 values for the test set compounds. 43,59 The values for the Q 2 and RMSE were calculated using the following equations:…”

“…43,59 The predictive capability of these models for the training data was investigated by calculating their LOO-cross-validated R 2 (Q 2 ) values, whereas the external predictability was evaluated by the externally crossvalidated R 2 values for the test set compounds. 43,59 The values for the Q 2 and RMSE were calculated using the following equations:…”

“…The applicability domain (AD) analysis is one of the crucial steps to identify the outlier molecules for any QSAR study. 27,39,59 Here, in this study, the leverage and the studentized residual values…”

Ligand and Structure-Based In Silico Determination of the Most Promising SARS-CoV-2 nsp16-nsp10 2′-o-Methyltransferase Complex Inhibitors among 3009 FDA Approved Drugs

“…Finally, the development of structure–activity relationship (SAR) principles [ 13 ]. Consequently, computational chemistry methods were applied to estimate various pharmacodynamic and pharmacokinetic parameters that relate the chemical structure of compounds to its activity and also to characterize the interaction of compounds with biological targets such as structure similarity [ 14 ], molecular fingerprints [ 15 ], QSAR [ 16 ], pharmacophores [ 17 ], homology models [ 18 ], molecular modeling [ 19 ], drug molecular design [ 20 ], rational drug design [ 21 , 22 ], molecular docking [ 23 ], MD simulations [ 24 ], absorption [ 25 ], distribution [ 26 ], metabolism [ 27 ], excretion [ 28 ], and toxicity properties [ 29 ], as well as physicochemical characterization [ 30 ] and DFT [ 31 ].…”

QSAR and Drug-likeness Studies of Thiadiazole Derivatives Against Lung Cancer

“…For optimization compounds, density functional theory (DFT) is an accurate but time-consuming method. The Semi-empirical Hamiltonians method can generate valid molecular parameters for creating QSAR models in a more time-efficient manner, especially when there is a lack of experience with descriptor selection (11)(12)(13).…”