Abstract: Project 4, Duke University Leptospirosis is a common, potentially life-threatening infection of global significance. The bacteria spread from a variety of different host animals to humans typically through contact with infected urine. As a result, leptospirosis is a leading cause of fever in the tropics. Leptospirosis can cause severe disease resulting in organ failure and death. Although the infection is responsive to antibiotics, treatment is often delayed due to similarities with other causes of fever (e.g., dengue) and limitations of current diagnostic tests. Leptospira spp. do not grow in standard culture media, molecular pathogen detection methods suffer from insufficient sensitivity throughout the course of the disease, and acute-phase serology is both insensitive and non-specific. Paired (acute and convalescent) sera can confirm acute infection retrospectively, but convalescent follow-up is infrequent. A growing body of evidence supports the potential for the development of molecular biomarkers based on the human immune response to the infection to not only detect infectious pathogen but also predict disease severity in infectious diseases. The development and validation of these “host response” diagnostic tests requires rigorous prospective studies of leptospirosis in diverse locations where different strains of the bacteria are common in people of different ancestral backgrounds. Our research team has 20 years of experience conducting such studies on 3 continents: Sri Lanka (Galle), Nicaragua (Leon), and Tanzania (Moshi) where the epidemiology of leptospirosis is well established. In addition to being relatively common, severe disease is well documented in Nicaragua and Sri Lanka to support development of diagnostics predictive of disease progression. Our team has extensive experience in generating the relevant molecular data (gene and protein expression) and the analytical, machine- learning tools for developing multi-analyte host response classifiers for pathogen differentiation across the course of infection which can be applied to identify and predict the severity of leptospirosis. The team also has experience translating these sets of biomarkers onto existing diagnostic platforms. In this project, we will add to our existing biorepository of samples from three geographically diverse locations to support the development of leptospirosis-specific molecular classifiers for both diagnosis and prediction of disease severity. Importantly, critical preliminary data and an unparalleled biorepository of samples and clinical data support Project 4 and synergize with Projects 1-3 in the proposed research program. Project 4 will be led by Megan E. Reller, MD, PhD, MPH.