EXPLORATION OF QSAR-BASED VIRTUAL SCREENING FOR THE DISCOVERY OF QUINOLONE-BASED ANTIBACTERIAL DRUGS

Abdulhaqq Taiwo Abdulwaheed*, Toheeb Ayodeji Sodiq, Wahab Abolore Badmus, Makuochukwu Collins Mbah, Olaitan Ebenezer Oluwadare, Miracle Oladoyin Ojedayo, Emmanuel Faderin, Precious Ifeanyi Ebere and Oluwagbade Joseph Odimayo

ABSTRACT

Drugs discovering is essential for assessing the potential impact on human health. Using 2D autocorrelation descriptors as predictor variables, a binary logistic regression model was developed to identify active antibacterial among quinolone compounds. The classifications made by the model on the training set compounds resulted in an overall accuracy, sensitivity and specificity of 91.80%, 90.62%, 93.10% dataset. The areas under the ROC curves, constructed with the training set data, was found to be 0.933 for the model. Predictions made by the model on the dataset to the test sets correctly classified 93% of test set compounds selected from datasets. The developed models are considered reliable for rapid discovery of drugs.

Keywords: Autocorrelation descriptor, Binary logistic regression, antibacterial, Quantitative structure-activity relationship.