Prélèvement dans l’espace de tête de composés responsables de l’odeur de verrat pour une détection à partir de composés organiques volatils

Given increasing animal welfare concerns on surgical castration of piglets without pain reliefs, several alternatives have been suggested. Amongst these are found, immunocastration – which can present some non-responders, and raising of entire males. In both cases (although significantly smaller in the first one), there is an increasing risk of boar taint in pork meat. The latter is a strong unpleasant smell and taste present when cooking or consuming pork, which is due to the release of several molecules present in the fat. Androstenone and skatole are the two compounds known to be the major responsible of this smell, giving respectively a strong urinal and fecal smell to the meat. Tainted carcasses must therefore be distinguished from untainted carcasses when being slaughtered. The present thesis has therefore focused on the detection of boar taint based on volatile organic compounds (VOCs). This research was performed as part of the AGROSENSOR project which aims at developing VOC-based sensors (gathered to function as an electronic nose) for the detection of non-compliant food matrices, such as tainted pork meat. Detection of boar tainted carcasses is often performed in slaughterhouses through an on-line sensory evaluation (i.e. human nose method). Although assessors can be rigorously trained prior to actual evaluation, this one is still subject to several factors impacting overall assessment of the taint (leading for example to a poor intra-rater reproducibility). Research on alternative methods for fast detection of boar taint has been on-going for decades already. The first part of this thesis therefore aims at reviewing the evolution of research in alternative detection techniques while also suggesting leads for future research. Some currently investigated techniques, such as laser diode thermal desorption combined to tandem mass spectrometry (LDTD- MS/MS), have proved their efficiency and are suggested for use in slaughterhouses. However, given elevated investment costs and the need for constant high slaughtering rates to reduced cost of analysis, other methods such as sensor-based methods were put forward for future research. A consensus was reached on the fact that current sensory evaluation will remain a method of choice in the future for many slaughterhouses. Sensory evaluation is the only method taking into account the total picture of boar taint (complete and complex assembly of various VOCs). Several discrepancies are therefore observed between classification based on quantification of skatole and androstenone and the human nose method. Understanding which other factors are responsible for such classification through sensory evaluation is unavoidable. This research has therefore linked scores attributed during sensory evaluation to chemical analyses of skatole and androstenone, fatty acid composition and other VOCs emitted along with skatole and androstenone when heating backfat. Taken together, the findings suggest that the score attributed during sensory evaluation is mainly due to boar taint compounds rather than differences in fat composition and general VOC emissions. Skatole and androstenone were emitted along with a large variety of oxidation products when fat was heated at elevated temperatures. Although this general observation was made with various heating temperatures, it appeared that the VOC profiles generated (which includes both boar taint compounds and non-boar taint compounds) were extremely dependent on the temperature used. This therefore stresses the urgent need for standardized human nose protocols. Additionally, low headspace concentrations in skatole and androstenone demonstrated that extraction conditions needed to be further optimized to allow VOC-based methods, such as sensor technology, to function within an acceptable range of concentrations. Further optimization of sampling parameters was performed. This included further increasing the heating temperatures while simultaneously adapting the sampling atmosphere. A combination of parameters was suggested, as a result, for boar taint compound extraction prior to detection through the use of analytical methods and also to complement current sensory analysis. Overall, this thesis was a steppingstone to unravel VOC profiles generated when heating pork fat and better comprehend the inherent difficulties associated to VOC-based detection of boar taint, sensory evaluation but also other VOC-based methods such as analytical methods used in laboratories and those being developed for on-line use in slaughterhouses.Agrosensor9. Industry, innovation and infrastructur