Quantitative Structure-Activity Relationship (QSAR) Study of a Series of Chalcone Derivatives Inhibiting Plasmodium Falciparum 3D7
Georges Stéphane Dembélé, Mamadou Guy-Richard Koné, Bafétigué Ouattara, Fandia Konaté, Doh Soro, Nahossé Ziao · American Journal of Physical Chemistry · 2022
This Quantitative Structure-Activity Relationship (QSAR) study was conducted using a series of twenty (20) chalcone derivatives with inhibitory activities against Plasmodium falciparum 3D7. The molecules were optimized at the B3LYP/LanL2DZ computational level, to obtain the molecular descriptors. This work was performed using the Linear Multiple Regression (LMR) method, the NonLinear Regression (NLMR) and the Artificial Neural Network (ANN) method. These tools allowed us to obtain three (3) quantitative models from the quantum descriptors that are, the overall softness (S), the bond lengths l(c=o) and l(c=c), and the polarizability (α). These models have good statistical performance. Among them, the ANN has a significantly better predictive ability R2 =0.997; RMCE = 0.035; F= 3571.499. The external validation tests verify all the criteria of Tropsha et al. and Roy et al. Also, the applicability domain of this model determined from the levers shows that a prediction of the pIC50 of new chalcone derivatives is acceptable when its lever value is lower than 1.07. For the ANN method, the Ch19 molecule is certainly outside the applicability domain, but it is not an influential point for the model, because this derivative belongs to the validation set, and therefore was not used in the model development. The behavior of this molecule could be explained by its structural diversity.