November 8, 2024 — A breakthrough in diabetes research has been unveiled today with the release of a flagship dataset from the AI-READI (Artificial Intelligence Ready and Equitable Atlas for Diabetes Insights) study. This NIH-backed initiative is pioneering the use of artificial intelligence to unlock complex interactions between biomarkers, environmental factors, and type 2 diabetes development. This release offers scientists globally access to data intended to revolutionize insights into both disease prevention and pathways to better health.
Unlike previous datasets, this new resource includes information on people both with and without type 2 diabetes at various stages, offering a more detailed perspective on the disease. Early findings are already surfacing novel associations. For instance, data from customized environmental sensors reveal that exposure to certain particulate pollutants is correlated with different stages of the disease. Additionally, the dataset includes survey responses, eye-imaging scans, depression scales, glucose measures, and other biological indicators.
“This level of granularity allows us to see that type 2 diabetes isn’t a one-size-fits-all condition. With this expansive data, we can explore the differences and subtleties among individuals,” said Dr. Cecilia Lee, professor of ophthalmology at the University of Washington School of Medicine and program director of AI-READI.
The initial dataset, published today in Nature Metabolism, comprises data from 1,067 people, representing just 25% of the anticipated 4,000 total participants. The researchers aim to collect a racially and ethnically diverse pool, divided evenly across categories by race/ethnicity, diabetes stage, and sex.
AI-READI, hosted by a consortium of seven institutions, brings together previously unaligned teams and disciplines for a common goal: leveraging unbiased, highly secure data to answer critical questions in diabetes. In addition to pathogenesis—the study of how diseases develop—Dr. Aaron Lee, the project’s principal investigator and a professor at UW Medicine, emphasized that the dataset also offers potential insights into salutogenesis, or factors promoting health.
“We’re not just asking what leads to disease,” Aaron Lee said. “We’re also interested in understanding what drives people toward health.”
The flagship dataset is hosted on a dedicated online platform in two formats: a controlled-access version, which requires user registration and a usage agreement, and a public, HIPAA-compliant version. Since the pilot data release earlier this year, over 110 research institutions worldwide have downloaded it, marking a high demand for AI-driven diabetes insights.
The AI-READI Consortium includes the University of Washington School of Medicine, University of Alabama at Birmingham, University of California San Diego, Johns Hopkins University, Native Biodata Consortium, Stanford University, and Oregon Health & Science University, with a central base at UW Medicine’s Angie Karalis Johnson Retina Center.
The hope is that this dataset will enable the creation of comprehensive health histories, predicting disease progression as well as pathways toward health—marking a promising shift in diabetes research from disease treatment to proactive health.