Description
Any food chain can be affected by food frauds, from mislabelling to authentication counterfeiting. The meat chain is no exception: in the case of processed meat, especially when it is sold minced or in sausage form, it becomes very difficult to distinguish the different ingredients, therefore mislabelling and meat substitution can easily occur. This is especially true with the famous German bratwurst, a type of sausage produced from minced or mechanically separated meat (MSM). MSM is obtained through a high-pressure process aimed at separating the bone from the edible meat tissue: the resulting pureed material is then formed into sausages and cooked. MSM meat is less expensive and of lesser quality compared to selected meat cuts, thus providing an economic incentive for the substitution fraud.
The present study was developed with the aim of determining whether NIR spectroscopy could be used for identifying meat products containing MSM, by means of chemometrics modelling. Alongside the main classification aim, a parallel research line regarding the actual quantification of MSM was developed and is the subject of this study. Bratwursts containing different percentages of MSM were minced and mixed with meat from non-MSM products, with the aim of obtaining new samples with specific MSM percentages. A calibration set of 30 samples spanning the content range between 0 % and 91% (in steps of 10 %) was built, together with a set of 27 samples corresponding to the “5 %” percentages (also in steps of 10 %). All samples were measured using three NIR instruments: the benchtop MPA (Bruker), the portable MicroNIR (Viavi) and the handheld SCiO sensor (Consumer Physics).
One PLS regression (Wold et al., 2001) model for each NIR dataset was developed and validated, and all three analytical techniques yielded good performances in calibration and prediction, with R2 values above 0.95.
