The acceleration of translational research requires new strategies, methods, and bioinformatic tools to predict and validate promising candidate therapies prior to investing nearly $1 billion and 5-10 years on clinical trials. One such strategy is to identify new therapeutic uses of existing drugs - also known as drug repurposing. However, no systematic process exists to address this challenging problem. We present a novel evidence- based computational framework to integrate the vast amount of scientific and health-related data sets with anonymous electronic health records, called Serendipitous Therapy Discovery. This framework provides an important bridge between basic science and clinical application. We will create the Serendipitous Therapy Discovery framework in three logical steps corresponding to the three specific aims. First, we will identify a large number of candidate therapies using the standard network-based approach of linking drug-target, drug- disease, and disease-gene interactions. We will integrate these networks to create a holistic view of drug- target-disease-gene that will identify new uses for existing drugs. Second, we will screen the candidate therapies against the FDA's Adverse Event Reporting System (AERS) and the Canada Vigilance Online Database (more than 4 million combined reports over nearly 45 years). This comprehensive compendium of pharmacovigilance data will enable us to identify and remove those therapies with serious adverse events. In addition, we will incorporate the Side Effect Resource (SIDER) database to identify those therapies with negative side effects. Third, we will validate the filtered candidate therapies using Electronic Health Records (EHRs) from Beth Israel Deaconess Medical Center (~3 million) and the Centers for Medicare &Medicaid Services (~12 million). As a first pass, we will identify any patients who are serendipitously on the therapy associated with the condition to evaluate if this therapy is safe in humans. Next, we will evaluate the effectiveness of repurposing the therapy by comparing the overall survival between patients on the candidate therapy versus those on standard therapy for that condition. In summary, Serendipitous Therapy Discovery is designed to recursively interrogate currently available scientific, health, and electronic medical data to provide evidence supporting the new use of a drug or drug combination.

Public Health Relevance

To accelerate bench-to-bedside drug discovery, we will develop a novel evidence-based computational framework to systematically identify new uses for existing drugs. We will do this using multiple public data sources to first, identifying a large number of candidate therapies, followed by screening out undesirable candidates, and finally validating the candidate therapies using anonymous medical records.

National Institute of Health (NIH)
National Library of Medicine (NLM)
Career Transition Award (K99)
Project #
Application #
Study Section
Special Emphasis Panel (ZLM1-ZH-C (J2))
Program Officer
Sim, Hua-Chuan
Project Start
Project End
Budget Start
Budget End
Support Year
Fiscal Year
Total Cost
Indirect Cost
Harvard University
Schools of Medicine
United States
Zip Code
Fusaro, Vincent A; Daniels, Jena; Duda, Marlena et al. (2014) The potential of accelerating early detection of autism through content analysis of YouTube videos. PLoS One 9:e93533
Fusaro, Vincent A; Patil, Prasad; Chi, Chih-Lin et al. (2013) A systems approach to designing effective clinical trials using simulations. Circulation 127:517-26