Machine Learning based Prediction of Anatomical Therapeutic Chemical (ATC) Class of Drug Like Molecule

Abstract

Drug discovery is costly and complex process with low success rate. Machine learning based computational method has been applied in different fields including drug discovery. Prediction of Anatomical Therapeutic Chemical (ATC) class of drug like molecule can provide therapeutic area or organ or system on which drug can be used. Computational prediction of ATC class of molecules only from structure (2D of 3D) can expedite the drug discovery process because no lab resources are required and even it can be done before the synthesis of any target molecule. In present research, Machine learning based method was developed using information of all small approved drug molecules. High accuracy of prediction in case of all 14 ATC class in independent test sets was achieved which warrants high reliability and usage of developed method.