Adverse Drug Reaction (ADR) is a serious health issue worldwide that increases the morbidity and mortality of individuals. Personalized medicine, involving targeted drug therapy for individuals based on the gene variants they inherit would prevent ADR. However, Pharmacogenomic testing is not widely available globally and it’s not cost-effective. NHS digital data (United Kingdom) indicates that the no of individuals on antidepressants has been doubled from 36 million (2008) to 70.9 million (2018), in the past decade (Lacobucci, 2019). One of the complications of antidepressants are adverse drug reactions (ADR), which are known to increase the morbidity and mortality and are often associated with increased hospitalization, increase in medical cost and less adherence to prescription (Sankhi et al 2020). All commercially available drugs undergo two metabolic process (Phase I and Phase II drug metabolism). There are several enzymes encoded by genes (for instance CYP gene family) which involve in this process, variants in these genes change the way the drug is processed. There are several GWAS (Genome Wide Association Studies) research studies, which has identified gene variants, through which individuals can be categorized into slow metabolizer, intermediate metabolizer, extensive metabolizer, and ultra-rapid metabolizer and drugs could be tailored, dose and the type of medication that could be prescribed based on an individual’s genotype. There are several pharmacogenomic studies that had identified potential genetic variants in CYP genes, that are associated with ADR in several diseases.
At this juncture we are proposing an interdisciplinary research project, adapting artificial intelligence, as a preliminary sciathon project, we would test the application of Artificial Intelligence (AI) in ADR prognosis, using antidepressants and its associated gene variants that causes adverse drug reactions to create a simulation platform, which could be used for the prognosis of adverse drug reactions to novel drugs in the market (in a similar way) and which has a potential application to create a black box warning about pharmacogenomic profiling for the usage of the drugs (this could be functionally validated by a separate research study).