2020/04/23 by Hind Lafridi, Lafridi, H., A. Oussa +5
Chemistry · Computer Science · Medicine · #2D-QSAR #Artificial neural network #Computational Drug Discovery Methods #DFT Study #Domain of Applicability. #Flavonoid derivatives #Free Radicals and Antioxidants #MLR #MNLR #PCA #Phytochemicals and Antioxidant Activities
paper · doi:10.48317/imist.prsm/morjchem-v8i4.20579
openalex publication_date 2020/04/23 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/01
This paper describes a quantitative structure – activity relationship study of the antiradical properties of 31 flavonoids belonging to different groups such as: flavonols, flavonones, dihydroflavonols and biflavonones. Using density functional theory (DFT) calculations, some structural characteristics such as frontier molecular orbitals, molecular descriptors, have been studied. To gain insights into the chemical structure and property of the studied compounds, many types of descriptors are generated by using DFT/B3LYP 6-31G(d,p) and other software. Also, The Principle Component Analysis (PCA), Multiple Linear and Nonlinear Regression (MLR and MNLR), and Artificial Neural Network (ANN) have been investigated to select the descriptors, and to generate the correlation models that relate the structural feature to the biological activity. The statistical results of the MLR, MNLR, and ANN indicate that the determination coefficient R2 were 0.811, 0.646, 0.982, respectively. A good correlation coefficient is obtained, and the antiradical activities of these compounds are well predicted. These models are expected to be useful for screening of polyphenolic antioxidants.