Virtual screening of drug candidate

Nanjing University Of Posts And Telecommunications | China

Wenyong Zhu | Haoran Liao | Hao Gao | Junyi Liang

        With the rapid development of life sciences at the end of the last century, the completion of the Human Genome Project (HGP) and the implementation of the follow-up functional genomics program, the pharmaceutical research and medical drug industry have entered a new historical period.         As we all know, the development of new drugs is a high-input, low-efficiency work. According to statistics, a new drug takes an average of 10 to 14 years from screening to successful listing, and the cost of the intermediate process is as high as 2 to 350 million US dollars. At present, the drug elimination rate in the clinical stage is as high as 90%. Therefore, how to shorten the time spent on the discovery and optimization of lead compounds and accelerate the speed of clinical research has become the focus of research by major pharmaceutical companies and academic institutions.         The main research involved in drug research is G Protein-Coupled Receptors (GPCR), which is an important protein for cell transmembrane signaling. It plays an important role in many important physiological activities of humans through its interaction with G proteins. GPCR is closely related to many diseases in humans, such as cancer and diabetes, and is also the target of about 40% of drugs on the market.         This project designed a new two-stage algorithm based on deep learning, WDL-RF, including molecular fingerprint generation stage based on new weighted deep learning and biological activity prediction stage based on random forest model, and applied to human important GPCR drugs. Finger molecule fingerprinting of the target (no three-dimensional structure) and prediction of biological activity values.