N. N. El Mrabiti, K. Dguigui , M. EL Hallaoui
A series of benzimidazole derivatives evaluated for their activities against four types of enteroviruses, coxsackievirus A16, B3, B6 and enterovirus 71 in Vero cells have been built and subjected to 3D-QSAR analysis to derive a correlation between TC50 and IC50 activities, and various physic-chemical descriptors. Both QSAR models were performed using multiple linear regression (MLR) and neural network (NN). Best correlations are obtained for both IC50 (rRLM = 0.90, rNN = 0.93), and TC50 (rRLM = 0.88, rNN = 0.95). To test the performance of these models, we used crossvalidation method with LOO procedure