MALDI-TOF mass spectrometry profiling of bovine skim milk for subclinical mastitis detection.

Based on signal intensity, a total of 24 peaks emerged as significant different between the two groups. The most discriminant signals (4,218.2 and 4,342.98 m/z) presented a ROC curve with AUC values higher than 0.8. Classification algorithms (i.e., quick classifier, genetic algorithm, and supervised neural network) were applied for generating models able to classify new spectra (i.e., samples) into the two classes. Our results support the MALDI-TOF mass spectrometry profiling as a tool to detect mastitic milk samples and to potentially discover biomarkers of the disease. Thanks to its rapidity and low-cost, such method could be associated with the SCC measurement for the early diagnosis of subclinical mastitis.