ARPA-H awards AI-driven project to repurpose approved medications


Biden Harris Administration’s ARPA-H awards AI-driven project to deliver progress for Americans facing a rare disease 

Initiative aims to accelerate drug discovery and improve patient outcomes by repurposing approved medications to treat rare and currently untreatable diseases 

Today, the Advanced Research Projects Agency for Health (ARPA-H) announced a new project to deliver hope and improved outcomes to those facing rare diseases, the ML/AI-Aided Therapeutic Repurposing In eXtended uses (MATRIX) project. MATRIX intends to build a machine learning platform to rapidly pinpoint and validate existing medications to treat diseases that currently have no therapies. For the more than 10,000 known rare diseases, today only a few hundred have safe, effective treatments. This project is in support of President Biden’s commitment to accelerate the development of cures for rare diseases and support medical research and innovation.   

“America’s health outcomes demand new innovation and action, which is exactly why President Biden started ARPA-H,” said the Director of the White House Office of Science of Technology Policy Arati Prabhakar, Ph.D. “This project will harness AI to find answers and treatments for the 30 million Americans who are currently facing rare diseases.”

“Rather than the current, one-step-at-a-time drug discovery process, we have an opportunity to use artificial intelligence to rapidly understand how already approved drugs could be effective against other diseases,” said ARPA-H Director Renee Wegrzyn, Ph.D. “Through this project, we hope to unlock the full potential of existing medicines to quickly and safely bring therapies to people with rare or currently untreatable diseases.”  

While a rare disease is defined as impacting less than 200,000 people in the United States, in total, these conditions impact 25 to 30 million, or one in ten Americans. If successful, MATRIX will develop an open-source platform and interactive heatmap displaying predicted efficacy scores for approved drugs against other diseases, thereby greatly accelerating traditional drug discovery timelines and providing an avenue to scale discovery across a vast set of diagnoses. ARPA-H is allocating up to $48 million to the non-profit Every Cure in Philadelphia in support of this research, with potential for additional funds to validate the most promising drug-disease matches in clinical trials. 

Every Cure will leverage ML/AI standards, knowledge graphs, and data models developed through the Biomedical Data Translator Program funded by the National Center for Advancing Translational Sciences, part of the National Institutes of Health.   

Wegrzyn added, “It’s exciting to consider the potential of a systematic approach to identifying and validating the top repurposing candidates across all drug-disease combinations. The team will work towards developing new treatment options and improved health outcomes for more people, all within significantly reduced timelines.” 

This project is one of many solicited through ARPA-H's Open Broad Agency Announcement (Open BAA). The Open BAA seeks transformative ideas for health research or technology breakthroughs. Continued support of each award is contingent on projects meeting aggressive milestones. To learn more about projects as they are awarded, visit the Open BAA page