Bacterial infections affect the lives of millions of individuals worldwide. When a patient is diagnosed with an infection, the exact cause is often unknown. Microbial diagnostics rely on methods requiring the growth of a bacterial colony, including lengthy and time-consuming culturing techniques and error-prone manual handling steps. Consequently, an accurate diagnosis may only be made after many days, by which time the patient could have received potentially unnecessary or non-effective antibiotic treatment. In the dairy industry, speeding up this diagnostic process would facilitate considerable improvements in antibiotic stewardship.

This thesis describes the development of three methods developed for the rapid diagnosis of different bacterial species in bovine mastitis directly from samples requiring different levels of laboratory equipment, costs and expertise. Laser-assisted rapid evaporative mass spectrometry was used to identify molecular biomarkers in mastitic milk. Diagnostic performance and individual markers were further investigated using transcriptomic studies. The second method used conserved species-specific primers designed from response regulators which were combined with a portable PCR system and a diagnostic dipstick for the rapid detection of species specific DNA. For the third method a whole-cell bacterial biosensor was constructed for different mastitis pathogens, allowing for the detection of quorum sensing signals released by the bacteria without further equipment.

All three diagnostic methods were successfully established and tested for the rapid identification of major bacterial species in bovine milk. By combining metabolomic and transcriptomic data, I was able to show that the detected mass spectrometry signals originated from the bacteria as well as the host. All three methods can detect bacteria in less than 30 to 240 min, with more technologically complex methods and less manual handling, yielding higher information content. These methods allow for a diagnosis within the time needed to inform the treatment of the patient, which could help reduce the prescription of unnecessary antibiotics. The general principles of these techniques can be utilised for other pathogens and diseases, broadening the impact and applicability of rapid bacterial identification for antimicrobial resistance prevention.