Leveraging Big Data and Genomics to Accelerate Drug Discovery and Guide Personalized Patient Care

Adverse drug reactions are a leading cause of death in the United States. Recent advances in genomics have revealed that the root cause of many of these adverse events can be traced to genetic mutations and polymorphisms affecting specific drug targets. Take for example a patient who has a mutation that renders them insensitive to the antiplatelet effects of clopidogrel. Following a cardiac stenting procedure, they will remain vulnerable to developing clots in their stents despite receiving standard antiplatelet therapy. Conversely, the risk of bleeding may outweigh the benefits of clopidogrel therapy following a stenting procedure for a patient who carries a hypersensitive genotype. Therefore, providing clinicians with a patient’s genotype and associated drug sensitivities during the prescribing process can contribute to more informed decision-making and stands to improve patient care through personalized medicine.

At a population level, gathering large genomic datasets can help guide drug discovery programs. By identifying certain genetic loci that are highly associated with certain diseases, subsequent animal studies and clinical trials can focus on drugs that have known binding sites at those targets and their associated proteins. In this way, a drug that is currently used to treat one disease by binding to a known molecular target may be identified as a potential candidate to treat several other diseases that are associated with mutations at that same molecular target. This “in silico” strategy can dramatically accelerate the drug discovery process, which currently relies on an “educated guess” strategy to initially identify drug targets. Vanderbilt University Medical Center has been working closely with Epic to integrate these strategies into their electronic health record, combining genomics with vast de-identified clinical databases. This has also allowed them to conduct large clinical trials at a substantially reduced economic and time cost.

With the construction of vast databases well underway, the next big step will be to develop a pragmatic strategy to bring this information to the bedside in a way that can actually benefit patients. Vanderbilt have started implementing the strategy outlined above with the example of clopidogrel to alert physicians when they are ordering certain drugs for patients that have a high risk of developing adverse reactions based on their clinical and genomic data. Importantly, these alerts do not stop physicians from prescribing the drugs. Instead, their purpose is to help guide and contribute to informed decision-making. Initial data has revealed that approximately 60% of physicians who received these alerts ultimately decided to prescribe a different drug. Clinician education programs are currently being developed to optimize the effect of these alerts on management decisions and patient care. Currently, the system is limited to five main drugs, which are most often prescribed in the context of cardiology and immune modulation. However, the introduction of CYP2D6 into their genotyping program promises to expand the scope of this endeavor to drugs commonly administered in the perioperative period such as beta-blockers and analgesics. Ultimately, they aim to share these personalized medicine initiatives with other institutions through integration with their native electronic health record systems.

*Coverage of the T.H. Seldon Memorial Lecture: Personalizing Healthcare in the Era of Big Data given by Jeffrey Balser MD, PhD, Dean, Vanderbilt University School of Medicine; President and CEO, Vanderbilt University Medical Center