Breath Biomarkers of Disease and Health Outcomes (BREATHE OUT)

Millions of Americans live with metabolic diseases like type 2 diabetes, and many more are unaware they have it or are at risk for it. Screenings to identify metabolic diseases are commonplace in hospital settings, but require bloodwork, which can be inconvenient, uncomfortable, and hard for some people to access (e.g., rural communities). As a result, many people go undiagnosed for years and miss out on the benefits for early treatment that could protect their long-term health. Interestingly, every breath we exhale carries information about what is happening inside our bodies, much like how a breathalyzer can detect alcohol. This makes breath a promising, noninvasive way to check on someone's metabolic health.

In this COBRE project, we will study whether a person's breath can reveal how healthy their metabolism is, comparing people in great metabolic health, people at risk for metabolic disease, and people already living with type 2 diabetes. We will also test whether breath can predict metabolic health as well as, or even better than, the blood tests doctors use today. This work builds on our team's earlier discovery of a breath pattern that was related to clinical malnutrition, a finding we have patented related to the clinical screening process. Working with the COBRE's data science experts, we will use advanced machine learning approaches (artificial intelligence or AI) to sort through the thousands of molecules in each breath sample and pinpoint the ones that matter most for metabolic disease. The long-term goal is to create a simple, non-invasive breath test that could one day help catch metabolic disease early, anywhere from a doctor's office, to community screening events, to your own home.