Traditional and Machine Learning Approaches in Structure-Based Drug Virtual Screening

Computer-aided drug discovery (CADD) and artificial intelligence-driven drug design (AIDD) represent highly efficient strategies aimed at decrease time and economic expenditures in the pharmaceutical industry, and the representative approaches include virtual screening, prediction of protein-ligand interaction and drug pharmacokinetic properties, and drug design. Generally, virtual screening is the initial step in drug discovery, with the primary objective of identifying and generating potential candidates for lead compounds. In the past decades, several traditional and machine-learning based methods have been developed to improve the accuracy and speed of virtual screening. This review discusses the development of advanced structure-based virtual screening methods by both traditional and machine learning approaches, including their performance, strength and limitations.