Mol Divers. 2022 Jul 17. doi: 10.1007/s11030-022-10495-5. Online ahead of print.
ABSTRACT
Human African trypanosomiasis (HAT) or sleeping sickness is a protozoan neglected tropical disease, which is the main health worry in more than 20 countries in Africa. A novel approach is presented to predict the antitrypanosomal activity of sesquiterpene lactones (STLs) in terms of biological activity (pIC50). The largest reported data set of pIC50 for Trypanosoma brucei rhodesiense (Tbr) as one form of HAT are used to derive and test the new model. The new model is based on five additive and two non-additive molecular structural parameters in several frameworks where it can be easily applied through a computer code. It is derived and tested based on 125 and 31 experimental data, respectively, with different types of statistical parameters. The high reliability of the novel model is compared with the best available QSAR models, which use “classical” molecular descriptors, and 3D pharmacophore features. The values of R2 (correlation coefficient), root mean squared error (RMSE), and RMSEP (root mean square error of prediction) of the new model are 0.77, 0.38, and 0.35, respectively. Meanwhile, R2, RMSE, and RMSEP of comparative QSAR models based on complex descriptors are in the ranges 0.71-76, 0.46-0.4, and 0.51-0.44, respectively. The predictive results of the novel approach confirm its high simplicity, reliability, precision, accuracy, and goodness-of-fit.
PMID:35842884 | DOI:10.1007/s11030-022-10495-5
