Drug Repurposing Using AI: Case Studies

Abstract

Drug repurposing, also called drug repositioning, is the process of identifying new therapeutics uses for existing drugs. A very full-cost and time-given undertaking of bringing in a whole new drug, it is mainly to this group of existing- approved and investigated drugs, where the researcher hopes to find something useful for a different disease or condition from that originally intended in the first place. In this study, three case studies are introduced which document the usage of AI as a tool in identifying new therapeutic indications for existing drugs. The first case details the invasion of the AI-generated pipeline to repurpose metformin-first-line therapy for type 2 diabetes-for Normal Tension Glaucoma (NTG) under the framework for Mendelian Randomization (MR). The study hypothesized its neuroprotection via glucagon and GLP1 pathways with some epidemiological evidence. The second case illustrates metformin as a supposed drug to reduce cancer mortality employing a range of large electronic health records (EHR) databases, which were merged with natural language processing and statistical modeling. The analysis showed statistically significant improved survival in metformin users in many cohorts. The third one indicates the repurposing of drugs developed for Fragile X Syndrome (FXS) through high-throughput testing and machine learning. The DREAM-RD identified Sulindac and other agents as worthy candidates, some of which have progressed into full- blown clinical trials. All three cases come with limitations that pertain to either specificity to a population, sample size, or, even, too little clinical data for an efficacious recommendation; all three nevertheless begin to set a benchmark for AI-based drug repositioning. The examples furnish a template for fast-tracking therapeutic discovery and development, especially in the domain of poorly served or rarer diseases.