Pneumonia is a severe infectious disease and a leading cause of hospitalization and death. It exacts an enormous cost in economic and human terms. Timely diagnosis is crucial for the outcome of therapy and health care associated costs. However, the high morbidity and mortality of pneumonia is partly due to the lack of efficient diagnostics. Animals with an excellent sense of smell (e.g. dogs) are capable of identifying bacterial infections by sniffing out sputum samples or even the air surrounding the infected patient. We thus hypothesize that pneumonia caused by bacterial agents can be identified by analyzing characteristic volatile metabolites produced during infection in the lung, which are eventually exhaled in breath. The goal of this project is to rapidly (within 15 min) diagnose bacterial pneumonia using a breath test. Additionally, the test should be capable to identify a subset of responsible pathogens. We will deploy an innovative mass spectrometric breath analysis technique (secondary electrospray ionization-mass spectrometry; SESI-MS) in a clinical setting. SESI-MS combines real-time response and unparalleled sensitivity (parts-per-trillion) and selectivity (~300 compounds per breath sample). We hypothesize that such a comprehensive metabolic fingerprint will provide an accurate pathogen-specific signature. We will combine cutting-edge analytical techniques, bioinformatics, mice models and access to well characterized patients in a multidisciplinary approach to elucidate and quantify exhaled metabolites that are indicators of pneumonia. This will significantly improve the current clinical and epidemiological situation by: i) enabling pathogen-based antibiotic treatment (thus reducing antibiotic resistance); ii) improving patients’ outcome; iii) saving costs by supporting evidence-based hospitalization/outpatient decisions and iv) identifying altered metabolic routes, thus gaining insights on the mechanisms by which pathogens circumvent the human immune system.