Data-driven drug discovery for drug repurposing

Abstract

To improve the decreased efficiency of drug discovery and development, drug repurposing (also called drug repositioning) has been expected, that it is a strategy for identifying new medical indications for approved, investigational or suspended drugs. Particularly, according to the rapid expansion of medical and life science data and the remarkable technological progress of AI technology in recent years, the approach of computational drug repurposing has been attracted as one of the applications in data-driven drug discovery. Computational drug repurposing is a method of systematical and strategical research for identifying novel indication candidates and prioritizing the indication candidates based on the various profiles of drugs, genes, and diseases. In this review article, the typical data science techniques for data-driven drug repurposing, 1. drug-target interaction prediction, 2. transcriptomics-based approach by using differentially gene expression profiles, 3. natural language processing and word embedding, and their current status were summarized. We have also introduced a use case of data-driven drug repurposing for the PPARγ/α agonist Netoglitazone that we actually analyzed. In addition, as an excellent successful case of data-driven drug repurposing in recent years, we have also discussed a repurposing case reported by BenevolentAI in 2020, that Baricitinib has been identified as a potential intervention for COVID-19, based on immunomodulatory treatment by its mechanism of action as a JAK1 and JAK2 inhibition.