Abstract
The quantitative activity relationships (QSARs) of the radical scavenging activities of 96 natural antioxidants based on the graph of atomic orbitals by CORAL software. The QSAR models are developed with the representation of the molecular structure by the Simplified Molecular Input-Line Entry System (SMILES) and the Monte Carlo optimization was used for calculation of the correlation weights of optimal SMILES based descriptors. The data set was randomly distributed three times into the training, invisible training, calibration and validation set was examined. The R2 values of the validation set for splits 1 to 3 were 0.966, 0.921, and 0.886 respectively. These models are established using only SMILES information, without any information on physicochemical parameters, the quantum mechanics descriptors, or 3D descriptors of the antioxidant structures. The obtained results indicated that the proposed QSAR models can be successfully used for predictions of the radical scavenging activities of new natural and synthetic antioxidants.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have