The Research Progress of AI-assisted New Drug Screening Technology

Abstract

Drug R&D is a systematic project with high investment, high risk and long period, and the traditional methods are faced with double pressures of cost and time. Computer-aided drug design (CADD) has introduced four steps: target discovery, confirmation, lead compound discovery and optimization, but the prediction accuracy is low and the consideration of drug formation is insufficient. With the development of artificial intelligence and big data, AI shows its potential in drug discovery, and improves the accuracy and efficiency of virtual screening through machine learning and deep learning. AI-driven virtual screening technology accurately captures molecular characteristics, enhances the enrichment of active molecules, reduces invalid tests and saves research and development costs. At present, AI-assisted virtual screening combines the advantages of traditional computing and AI learning, and combines ligand-based, structure-based and AI-driven technologies to provide innovative solutions for drug discovery. Technical breakthroughs such as deep generation model indicate that the field of drug discovery will face changes. It is of great value to study the application of AI-driven virtual screening technology to improve research and development efficiency and reduce costs.