The KNIME based classification models for yellow fever virus inhibition

Abstract

Yellow fever is one of the virus-infected diseases spread through mosquitoes and kills more than thirty thousand people every year. Although a large number of compounds have been reported, none of the drugs have yet been approved for clinical use. In the process of drug development against yellow fever virus (YFV), in the present investigation, we have developed efficient classification models based on a large dataset (309 compounds) compiled from the ChEMBL database. The Naïve Bayes method as implemented in the KNIME platform was used for the classification analysis. The best models obtained using the combined dataset show accuracy of >90% on the test set prediction (Matthew's correlation coefficients of >0.7). All the models developed in this study could be applicable for virtual screening of yellow fever virus inhibition.