Metabolomics: bacteria in milk

 PhD-fellow: Kasper Hettinga (graduated 2009)

Abstract PhD thesis

The objectives of the work described in this thesis were twofold. The first aim was to test the suitability of headspace analysis for quality control of raw milk in general. The second aim was to further develop the headspace analysis for the identification of mastitis causing pathogens in raw milk. Fresh raw milk without quality defects was shown to always contain the same 7 volatile components. Treatments like heating and homogenization of raw milk drastically changed this basic pattern. Using the headspace analysis, variation in the composition of the regular diet could not be detected, and Pseudomonas only when present in high numbers. On the other hand, the headspace analysis can be used for quantifying the extent of lipolysis, the amount of chloroform, and the detection of feeding specific vegetable byproducts. The evaluated headspace method is thus able to detect several quality defects simultaneously and therefore is a valuable supplementary method for raw milk quality control. Additionally, it was investigated if headspace analysis could be used as a faster method for identification of mastitis causing pathogens, based on the analysis of volatile bacterial metabolites. This method, supported by an artificial neural network, was found to reliably identify the most important groups of mastitis causing pathogens. To study the origin of these metabolites, milk was inoculated with isolated mastitis pathogens. Most metabolites found in inoculated milk samples corresponded with their occurrence in mastitis milk. Finally, the influence of incubation on the formation of these metabolites was studied. Incubation was shown to be a  necessary step for the detection of volatile metabolites. It was found that after 8 hours of incubation, all important metabolites were formed.

Current research

The method described in the PhD thesis focused on identifying single mastitis pathogens. In the follow-up research three questions are addressed:
  1. Can other bacteria (e.g. butyric acid bacteria spores) be identified as well?
  2. What happens if multiple bacterial species are present in one sample?
  3. Can we use H-NMR to further characterize the metabolites formed by the bacteria?