Derivation of Highly Predictive 3D-QSAR Models for hERG Channel Blockers Based on the Quantum Artificial Neural Network Algorithm

• DOI: 10.3390/ph16111509
• Publication Date: 2023-10-24T10:28:25Z
• Authors: Taeho Kim; Kee-Choo Chung; Hwangseo Park
• Keywords: hERG

The hERG potassium channel serves as an annexed target for drug discovery because the associated off-target inhibitory activity may cause serious cardiotoxicity. Quantitative structure-activity relationship (QSAR) models were developed to predict activities against channel, utilizing three-dimensional (3D) distribution of quantum mechanical electrostatic potential (ESP) molecular descriptor. To prepare optimal atomic coordinates dataset molecules, pairwise 3D structural alignments carried out in order cross correlation between template and other molecules be maximized. This alignment method stands from common atom-by-atom matching technique, it can handle structurally diverse effectively chemical derivatives that share identical scaffold. problem prevalent 3D-QSAR methods was ameliorated substantially by dividing into seven subsets, each which contained with similar weights. Using artificial neural network algorithm find functional ESP descriptors experimental activities, highly predictive derived all subsets extent squared coefficients exceeded 0.79. Given their simplicity model development strong predictability, this study are expected function effective virtual screening tool assessing cardiotoxicity candidate molecules.