Near Infrared Analysis as a First-Line Screening Technique for Identifying Animal Species in Rendered Animal by-Product Meals

Abstract

The only official method currently available for the identification of feed ingredients in mixtures and compound feedingstuffs is inspection by optical microscopy (OM). From October 2003, the EU adopted Regulation (EC) No. 1774/2002 governing animal by-products (ABPs), which seeks to address the possible risks inherent in recycling potential infectivity due to the absence of barriers within species and to exclude the cannibalism which may be induced by intra-species recycling. The main aim of the present work was to develop and validate near infrared (NIR) spectroscopy chemometric models for the identification of the animal species in ABPs. A total of 352 meat and bone meal (MBM) samples from different species ( n = 80 pure poultry MBM, 75 pure pork MBM and 197 mixtures of cattle MBM with other species) were scanned using a monochromator instrument equipped with a transport module. The total sample set was split into a training set ( n = 234) and a validation set ( n = 118). Various mathematical and scatter correction treatments were tested and used to develop two different PLS2 discriminant models: Model I, to discriminate between ruminant and non-ruminant meals; and Model II, to discriminate between poultry by-product meal, pork meal and a mixture of cattle meal with other species. All models yielded SECV values approaching 0.2 and over 95% correct classification in the training step; whereas in the validation step, approaching 90% (Model I and II) of samples were correctly classified. The methodology presented here, based on discriminant analysis of NIR spectra, for use as a “first line of defence” in identifying animal species in rendered ABP meals, provides a reliable, fast and affordable means of enforcing legislation concerning the ban on MBM.