Application of spectroscopic techniques and chemometric approaches for authentication of animal food products

With increasing complexity of food chains and hidden food fraud, there is a need to develop analytical methodologies to tackle authenticity concerns, reassure safety parameters and ensure product quality. This body of work is divided in four case studies (Chapters 3, 4, 5 & 6) regards different applications of NMR and NIR spectroscopic techniques to evaluate authenticity of agri-food products from animal origin. In Chapter 3, a bench-top and a portable NIR spectrometer are used to discriminate pyrrolizidine alkaloids (PAs) and their N-oxides (PANOs) in contaminated dehydrated bee pollen samples. The application of CDA resulted in a modelling statistical approach that demonstrates the predictive capacity of NIR systems to distinguish among the three quantitative PA/PANOs classes, especially for detection of those samples belonging to the low class, which corresponds to safe samples. In Chapter 4, the feasibility of using bench-top and portable NIR spectrometers combined with multivariate statistical models for discriminating chicken breast shelf life is explored. NIR spectroscopy showed reliable effectiveness to recognise a 7-day shelf life threshold of breasts, suitable for routine at-line application for screening of meat quality. In Chapter 5, capability of a VIS/NIR and two NIR instruments (portable and bench-top) to discriminate among table eggs from quails fed with different inclusion levels of silkworm pupa meal was evaluated. Both NIR benchtop and portable devices combined with PLS-DA, KNN and SVM successfully showed capacity to recognise in the eggs the inclusion of insect meal in layers diet while the VIS-NIR portable tool displayed worse predictive capacity. In Chapter 6, discriminant capacity of fatty acids and NMR metabolomic profiles of milk from three different forage-based dairy chains was tested. The outcomes showed that only a total replacement of maize silage with legume and grass hays in the cows’ diet led to a significant change in the milk metabolomic profile. A low-level FA and NMR data fusion coupled with a CDA chemometric approach has been shown to improve the predictive performance of the supervised CDA discriminant model of milk samples from diverse ensiled or dried forage-based feeding systems. Our findings suggest that both NIR and NMR spectroscopic techniques are effective methods for the authentication and analysis of the studied food products. Implementation of such methods in routine authentication analysis while keeping the costs low and the performance level high can open up the potential for further incorporation of NMR and NIR technologies in food analysis.