STRATFEED NIRS-NIRM Method

 

Near infrared (NIR) light reflected off feeds produces a spectral "signature" that is unique to that feedstuff. It behaves like a "finger-print". NIR light is absorbed by vibrating bonds in molecules that are characteristic of the chemical components of food like water, oil, starch, sugar, fibre and protein. Mammalian tissue such as banned meal and bone meal (MBM) has its own unique spectral "signature" such that it is possible to build mathematical models to detect the presence of MBM in feed by NIR. This method has the advantage to be independent of human subjectivity.

 

The discriminating power of a infrared spectrum is very high and no confusion is possible between a spectrum of soya and a spectrum of MBM. Nevertheless, the inconvenient of a macroscopic IR measure is the unaccuracy around very low concentrations. It's the raison why, beside the Near InfraRed Spectroscopy (NIRS), the Near InfraRed Microscopy (NIRM) has been investigated through the STRATFEED project to be able to analyse the sample at the particle level.

 

Some publications are referenced at the end of this document

 

The Near InfraRed Spectroscopy (NIRS)

 

The NIRS method consists in the macroscopic analysis of feedingstuffs samples by IR. With this technique one spectrum by sample is acquired. One of the main features of the NIRS method is the possibility to analyse a large quantity of specimen. Indeed, analysis is performed on 10 g to 100 g of sample, allowing the natural inhomogeneity of feed samples to be taken into account and reducing the probability of false-negatives resulting from inappropriate sampling procedures.

 

The STRATFEED project aims to provide protocol for spectra acquisition and instrument standardisation. Large spectral libraries regrouping hundreds of spectra of pure feedingstuffs and adulterated samples have been build to develop mathematical models. Results interpretation using the best models are showed through the STRATFEED explorer

 

In this project, a Foss NIRSystemsä model 6500 (Figure 1) was used. It takes only a minute to acquire the spectrum of a feed and to report its composition and thus to flag any suspect specimens and refer them for testing by more costly, time-consuming PCR analysis, for instance. At present, the major drawback of the NIRS technique is that the LOD is higher than 1% and the method cannot be used alone. Nevertheless, as NIR is widely used in the feed industry for routine quality control and control of feed materials, it may be possible to use NIR routinely to screen samples for MBM contamination. NIR may be the first line of defence of the food chain against material that risks the transmission of TSE's.

 

                                        Figure 1. :Foss NIRSystemsä model 6500.

 

 

The Near InfraRed Microscopy (NIRM)

 

In this project, an Auto Image Microscope connected to a Perkin-Elmer Fourier transform near infrared (FT-NIR) spectrometer was used (Figure 2). This instrument allows collecting spectra from small particles (50µ). This material works as follows: the beam generated by the interferometer goes to the microscope, which is equipped with a video camera that allows to visualize the sample and to localize the particles we want to analyze. The infrared rays are focused into particles to analyze and a detector located in the microscope measures the reflected beam. Then an inverse Fourier transform is applied. The great advantages of this technique are that the recognition is not dependent on the expertise of the analyst and that it is possible to automate all procedures and to analyse more samples per unit of time than the classical microscopy. The figure here below presents the FT-Near-infrared microscope used in this study. With this technique, spectra of hundred particles coming from one sample are analysed.

 

Figure 2. : Auto Image Microscope and Perkin-Elmer Fourier transform near-infrared spectrometer.

 

 

The STRATFEED project aims to provide protocol for detection of MBM by NIRM. Large spectral libraries regrouping hundreds of spectra of animal and vegetal particles have been build to develop mathematical models. Results interpretation using the best models are showed through the STRATFEED explorer

 

The detection of MBM by NIRM can be carried out in 2 steps:

 

 

The results obtained in the framework of the STRATFEED project demonstrate that this method has a limit of detection (LOD) inferior to 0.1 %  and allows the detection of MBM in fishmeal.

The data acquisition

 

The NIRM  method consists in the analysis of several hundreds of particles being the result of the grinding of a compound feedingstuffs. The analysis, as in classical microscopy, can be done on the raw material or on the sediment fraction. All the samples are provided by the STRATFEED European consortium (see European Sample Bank). The identification of the ingredients is made on the basis of spectral features observed in the near-infrared region (1100 - 2500 nm) of the electromagnetic spectrum. The sample particles are spread on a sampling holder and a spectrum of each particle is successively collected. A spectral library, on the Figure 3 here below, shows the sampling holder and the different spectra of several ingredients allowed and prohibited in the formulation of feedingstuffs. Note the spectral features differences observed for each source of ingredient.

 

 

Figure 3. : Sample particles spread on the sample holder and near-infrared spectra of different products.

 

The data treatment

 

All the spectral data collected are stored in a database. For data treatment, different statistical tools are available to analyse the spectral data and discriminate between the origins of the different particles making up a feedingstuff. Three methods are here presented: SLDA , PLS and ANN methods.

 

SLDA method

 

       Stepwise Linear Discriminant Analysis (SLDA) can be used to construct the different discriminant equations. The first step consists in the assignment of an arbitrary number to each particle spectrum according to its origin. Then the SLDA procedure is applied to distinguish different groups. This step allows the extraction of the most relevant wavelengths for the discrimination of groups and to define the sources that are easier to discriminate. The procedure is applied until each group corresponds to a single particle source. The Figure 4 here below shows the results of the SLDA. According to the first dimension, the samples from vegetables meals are located on the left side of the graph while the samples from animal origin are located at the right side of the graph. According to the second dimension, the samples from meat and bone meals are on the top while those from fish meals are on the bottom.

Figure 4. : Discrimination between vegetable, MBM and Fish meals using the SLDA model.

 

PLS method

 

       Partial Least Squares (PLS) is a supervised method currently used data treatment and based on a linear model (linear combination of absorbances at different wavelengths). The Figure 5 here below shows the results of the PLS. The horizontal line marks the threshold used to separate groups encoded as -1 for animal particles and 1 for vegetable particles. Predicted values below 0 are assigned to the first group and values above 0 are assigned to the second group. Vegetable and animal particles are clearly differentiated.

 

Figure 5. : Discrimination between vegetable and animal groups using the PLS model.

 

 

ANN method

 

       Artificial Neural Network is another supervised method based on a non linear approach. In this method, the predictive model is under the form of a network of interconnected small units that individually perform a simple task. Each unit, called a "neurone", processes input information and communicates the result to another neurone of the network. The Figure 6 here below shows the results of the ANN. The horizontal line marks the threshold used to separate groups encoded as -1 for animal particles and 1 for vegetable particles. Predicted values below 0 are assigned to the first group and values above 0 are assigned to the second group. In this case, the values predicted by ANN model are more accurate than by PLS model.

 

Figure 6. : Discrimination between vegetable and animal groups using the ANN model.

 

é

Publications

 

Journal article

 

Bellorini, S., Strathmann, S., Baeten, V., Fumière, O., Berben, G., Tirendi, S. & von Holst, C. (2005). Discriminating animal fats and their origins : assessing the potentials of Fourrier transform infrared spectroscopy, gas chromatography, immunoassay and polymerase chain reaction. Anal. Bioanal. Chem., 382, 1073-1083.

Perez Marin, D. & Garrido Varo, A. (2005). The microscopic detection of animal proteins in feeds. Biotechnology, Agronomy, Society and Environment, 9 (1), 3-9.

 

Gizzi, G., von Holst, C., Baeten, V., Berben, G. & van Raamsdonk, L. (2004). Determination of processed animal proteins, including meat and bone meal, in animal feed. Journal of AOAC International, 87 (6), 1334-1341.

 

Gizzi, G., van Raamsdonk, L., Baeten, V., Murray, I., Berben, G., Brambilla, G. & von Holst, C. (2003). Risk analysis of prion diseases in animal. An overview of tests for animal tissues in feeds applied in response to public health concerns regarding BSE. Revue Scientifique et Technique, Office International des Epizooties, 22 (1), 311-331.

 

Baeten, V. & Dardenne, P. (2001). The contribution of near infrared spectroscopy to the fight against the mad cow epidemic. NIR news, 12 (6), 12-13.

 

 

Book

 

Baeten, V., von Holst, C., Fissiaux, I., Michotte Renier, A., Murray, I. & Dardenne, P. (2005). The near infrared microscopic (NIRM) method: combination of the advantages of optical microscopy and near-infrared spectroscopy (WP5-NIRM). In: Strategies and methods to detect and quantify mammalian tissues in feedingstuffs, Brussels, European Commission, 14p.

Murray, I., Garrido Varo, A., Perez Marin, D., Guerrero, J.E., Baeten, V., Dardenne, P., Termes, S., Zegers, J. & Frankhuizen, R. (2005). Macroscopic near-infrared reflectance spectroscopy (WP5-NIRS). In: Strategies and methods to detect and quantify mammalian tissues in feedingstuffs, Brussels, European Commission, 27p.

 

Baeten, V., Michotte Renier, A., Sinnaeve, G., Garrido Varo, A. & Dardenne, P. (2004). Analysis of the sediment fraction of feed by Near-Infrared Microscopy (NIRM). In: Near Infrared Spectroscopy: Proceedings of the 11th International Conference, Davies, A.M.C. & Garrido-Varo, A. Norfolk-UK, NIR Publications, 663-666.

Bautista, J., Garrido Varo, A., Guerrero, J.E., Perez Marin, D. & Buron, T. (2004). NIRS for animal species identification in animal protein by-products: a viability study. In: Near Infrared Spectroscopy: Proceedings of the 11th International Conference, Davies, A.M.C. & Garrido-Varo, A. Norfolk-UK, NIR publications, 651-654.

Corbisier, P., Sinnaeve, G., Sindic, M., Baeten, V., Dardenne, P. & Deroanne, C. (2004). Optimisation of the NIR measurement of meat and meat products; valorisation of the data bases. In: Near Infrared Spectroscopy: Proceedings of the 11th International Conference, Davies, A.M.C. & Garrido-Varo, A. Norfolk-UK, NIR publications, 535-540.

de la Haba, J., Garrido Varo, A., Guerrero, J.E., Perez Marin, D. & Buron, T. (2004). NIRS for animal species identification in animal protein by-products: a viability study. In: Near Infrared Spectroscopy: Proceedings of the 11th International Conference, Davies, A.M.C. & Garrido-Varo, A. Norfolk-UK, NIR publications, 647-650.

Michotte Renier, A., Baeten, V., Sinnaeve, G., Fernandez Pierna, J.A. & Dardenne, P. (2004). The NIR camera: a new perspective for meat and bone meal detection in feedingstuffs. In: Near Infrared Spectroscopy: Proceedings of the 11th International Conference, Davies, A.M.C. & Garrido-Varo, A. Norfolk-UK, NIR publications, 1061-1065.

Murray, I., Perez Marin, D., Garrido Varo, A., Guerrero, J.E., Puigdomènech, A., Dardenne, P., Baeten, V. & Zegers, J. (2004). STRATFEED Near-infrared instrument network for detecting animal tissues in feedingstuffs. In: Near Infrared Spectroscopy: Proceedings of the 11th International Conference, Davies, A.M.C. & Garrido-Varo, A. Norfolk-UK, NIR publications, 291-294.

Perez-Marin, D., Garrido, A., Guerrero, J., Murray, I., Puigdomenech, A., Dardenne, P., Baeten, V. & Zegers, J. (2004). Detection and quantification of mammalian meat & bone meal in compound feedingstuffs by NIR Spectroscopy. In: Near Infrared Spectroscopy: Proceedings of the 11th International Conference, Davies, A.M.C. & Garrido-Varo, A. Norfolk-UK, NIR publications, 667-671.

Termes, S., Tauler, R. & Puigdomènech, A. (2004). Preliminary results in the determination of meat and bone meal. In: Near Infrared Spectroscopy: Proceedings of the 11th International Conference, Davies, A.M.C. & Garrido-Varo, A. Norfolk-UK, NIR publications, 659-662.

 

Conference proceedings

 

Baeten, V., Michotte Renier, A., Sinnaeve, G., Garrido Varo, A. & Dardenne, P. (2003). Analysis of the sediment fraction of feed by Near-Infrared Microscopy (NIRM). 11th International Conference on Near-Infrared Spectroscopy: NIRS 2003 - stretching the NIR spectrum to the limit, Cordoba - Spain, 6-11 April 2003, UCO, P 5.83.

Bautista, J., Garrido Varo, A., Guerrero, J.E., Perez Marin, D. & Buron, T. (2003). NIRS for animal species identification in animal protein by-products: a viability study. 11th International Conference on Near-Infrared Spectroscopy: NIRS 2003 - stretching the NIR spectrum to the limit, Cordoba - Spain, 6-11 April 2003, UCO, 1p.

de la Haba, J., Garrido Varo, A., Guerrero, J.E., Perez Marin, D. & Buron, T. (2003). NIRS for animal species identification in animal protein by-products: a viability study. 11th International Conference on Near-Infrared Spectroscopy: NIRS 2003 - stretching the NIR spectrum to the limit, Cordoba - Spain, 6-11 April 2003, UCO, 1p.

Michotte Renier, A., Baeten, V., Sinnaeve, G., Fernandez Pierna, J. & Dardenne, P. (2003). The NIR camera: a new perspective for meat and bone meal detection in feedingstuffs. 11th International Conference on Near-Infrared Spectroscopy: NIRS 2003 - stretching the NIR spectrum to the limit, Cordoba - Spain, 6-11 April 2003, UCO,

Murray, I., Perez-Marin, D., Garrido, A., Guerrero, J., Puigdomenech, A., Dardenne, P., Baeten, V. & Zegers, J. (2003). STRATFEED Near-infrared instrument network for detecting animal tissues in feedingstuffs. 11th International Conference on Near-Infrared Spectroscopy: NIRS 2003 - stretching the NIR spectrum to the limit, Cordoba - Spain, 6-11 April 2003, UCO, 4p.

Perez-Marin, D., Garrido, A., Guerrero, J., Murray, I., Puigdomenech, A., Dardenne, P., Baeten, V. & Zegers, J. (2003). Detection and quantification of mammalian meat & bone meal in compound feedingstuffs by NIR Spectroscopy. 11th International Conference on Near-Infrared Spectroscopy: NIRS 2003 - stretching the NIR spectrum to the limit, Cordoba - Spain, 6-11 April 2003, UCO, 5p.

Termes, S., Tauler, R. & Puigdomènech, A. (2003). Preliminary results in the determination of meat and bone meal. 11th International Conference on Near-Infrared Spectroscopy: NIRS 2003 - stretching the NIR spectrum to the limit, Cordoba - Spain, 6-11 April 2003, UCO, 1p.

 

Michotte Renier, A., Baeten, V. & Dardenne, P. (2002). Detection and quantification of meat and bone meal using a near infrared camera. 3rd International Chemometrics Research Meeting, The Netherlands, 26-30 May 2002, 1p.

Murray, I. (2002). Theory, principles and practice of analysis by NIR spectroscopy. NIRS Workshop 2002 : Application of NIRS Technology for Agricultural Commodities and Products, Edmonton, Alberta - Canada, 30 September - 4 October  2002, University of Alberta, 7p.

Murray, I. (2002). Discriminant and authentication analysis for feeds and foods. NIRS Workshop 2002 : Application of NIRS Technology for Agricultural Commodities and Products, Edmonton, Alberta - Canada, 30 September - 4 October  2002, University of Alberta, 8p.

Murray, I. (2002). What role does NIR spectroscopy offer for meat quality and safety research? NIRS Workshop 2002 : Application of NIRS Technology for Agricultural Commodities and Products, Edmonton, Alberta - Canada, 30 September - 4 October  2002, University of Alberta, 8p.

Murray, I. (2002).

Detection of adulteration of fishmeal & feeds with meat and bone meal (MBM) by discriminant analysis on visible and NIR Spectra.

Invited lecture presented at the NIR Spectroscopy meeting 14th Nov 2002, Tsukuba, Japan.

 

Baeten, V., Michotte Renier, A., Sinnaeve, G. & Dardenne, P. (2001). Analysis of feedingstuffs by near-infrared microscopy (NIRM) : detection and quantification of meat and bone meal (MBM). 6th International Symposium on Food Authenticity and Safety, Nantes - France, 28-30 November 2001, FASIS, 11p.

 

Posters

 

Baeten, V., von Holst, C., Michotte Renier, A., Fissiaux, I., Murray, I. & Dardenne, P. (2004). In-house validation of the near infrared microscopy (NIRM) technique for the detection of animal meal in feedingstuffs. Poster in: International symposium on Food and Feed safety in the context of prion diseases, Namur - Belgium, 16-18 June 2004.

Bautista, J., Garrido Varo, A., Guerrero, J.E. & Perez Marin, D. (2004). The use of NIR spectral signature to differentiate among categories of inedible fats and oils. Poster in: International symposium on Food and Feed safety in the context of prion diseases, Namur - Belgium, 16-18 June 2004.

de la Haba, J., Garrido Varo, A., Guerrero, J.E. & Perez Marin, D. (2004). Near Infrared Reflectance Spectroscopy (NIRS) for identification of the specie in animal protein processed by-products. Poster in: International symposium on Food and Feed safety in the context of prion diseases, Namur - Belgium, 16-18 June 2004.

de la Roza, B., Perez Marin, D., Garrido Varo, A. & Guerrero, J.E. (2004). Application of Support Vector Machine (SVM) as chemometric model on NIRS data base to classify compound feedingstuffs contaminated with animal meals. Poster in: International symposium on Food and Feed safety in the context of prion diseases, Namur - Belgium, 16-18 June 2004.

Fernandez Pierna, J.A., Michotte Renier, A., Baeten, V. & Dardenne, P. (2004). NIR camera and Chemometrics (SVM): the winner combination for the detection of MBM. Poster in: International symposium on Food and Feed safety in the context of prion diseases, Namur - Belgium, 16-18 June 2004.

Perez Marin, D., Garrido Varo, A., Guerrero, J.E. & de la roza, B. (2004). In-house validation of NIRS chemometric models to test the presence of animal origin meals in unground compound feedingstuffs. Poster in: International symposium on Food and Feed safety in the context of prion diseases, Namur - Belgium, 16-18 June 2004.

Perez Marin, D., Garrido Varo, A., Guerrero, J.E. & Termes, S. (2004). External validation of NIRS models developed in the framework of the STRATFEED European project to predict the percentage of animal origin meals in compound feeds. Poster in: International symposium on Food and Feed safety in the context of prion diseases, Namur - Belgium, 16-18 June 2004.

Termes, S., Bosch, J., Tauler, R. & Puigdomènech, A. (2004). Validation of the determination of meat and bone meal in feedingstuff with NIR spectroscopy. Poster in: International symposium on Food and Feed safety in the context of prion diseases, Namur - Belgium, 16-18 June 2004.

 

Baeten, V., Michotte Renier, A., Sinnaeve, G., Garrido Varo, A. & Dardenne, P. (2003). Analysis of the sediment fraction of feed by Near Infrared Microscopy (NIRM). Poster in: 11th International Conference on Near Infrared Spectroscopy, Cordoba - Spain, 6-11 April 2003.

Bautista, J., Garrido Varo, A., Guerrero, J.E., Perez Marin, D. & Buron, T. (2003). NIRS for animal species identification in animal protein by-products: a viability study. Poster in: NIR publications, Cordoba - Spain, 6-11 April 2003.

Corbisier, P., Sinnaeve, G., Sindic, M., Baeten, V., Dardenne, P. & Deroanne, C. (2003). Optimisation of the NIR measurement of meat and meat products; valorisation of the data bases. Poster in: 11th International Conference on Near Infrared Spectroscopy, Cordoba - Spain, 6-11 April 2003.

de la Haba, J., Garrido Varo, A., Guerrero, J.E., Perez Marin, D. & Buron, T. (2003). NIRS for animal species identification in animal protein by-products: a viability study. Poster in: NIR publications, Cordoba - Spain, 6-11 April 2003.

Michotte Renier, A., Baeten, V., Sinnaeve, G., Fernandez Pierna, J.A. & Dardenne, P. (2003). The NIR camera: a new perspective for meat and bone meal detection in feedingstuffs. Poster in: 11th International Conference on Near Infrared Spectroscopy, Cordoba - Spain, 6-11 April 2003.

Murray, I., Perez Marin, D., Garrido Varo, A., Guerrero, J.E., Puigdomenech, A., Dardenne, P., Baeten, V. & Zegers, J. (2003). STRATFEED near infrared instrument network for detecting animal tissues in feedingstuffs. Poster in: 11th International Conference on Near Infrared Spectroscopy, Cordoba - Spain, 6-11 April 2003.

Perez Marin, D., Garrido Varo, A., Guerrero, J.E., Murray, I., Puigdomenech, A., Dardenne, P., Baeten, V. & Zegers, J. (2003). Detection and quantification of mammalian meat & bone meal in compound feedingstuffs by near spectroscopy. Poster in: 11th International Conference on Near Infrared Spectroscopy, Cordoba - Spain, 6-11 April 2003.

Termes, S., Tauler, R. & Puigdomènech, A. (2003). Preliminary results in the determination of meat and bone meal. Poster in: NIR publications, Cordoba - Spain, 6-11 April 2003.


Contributions

This work was coordinated by CRA-W in collaboration with JRC-IRMM, AFSCA, FUSAGx, RIKILT, NUTRECO, SAC, ISS, UCO, LAGC, ALP, LUFA and PDIR

 

Walloon Agricultural Research Centre

CRA-W (B)

Dr Pierre Dardenne, Dr Vincent Baeten, Dr Gilbert Berben,

Dr R. Oger, Ir Philippe Vermeulen

Institute for Reference Materials and Measurements

European Commission - DG - Joint Research Centre

JRC-IRMM (B)

Dr Christoph von Holst

Federal Feed & Food Laboratory of Tervuren

AFSCA-FVLT(B)

Ir Jeroen Vancutsem

Gembloux Agricultural University

FUSAGx (B)

Prof Daniel Portetelle

Institute of Food Safety

RIKILT (NL)

Dr Jacob de Jong, Dr Leo van Raamsdonk, Dr Henk Aarts, Mr Rob Frankhuizen

Masterlab BV

NUTRECO (NL)

Ir Jos Zegers

Scottish Agricultural College

SAC (UK)

Dr Ian Murray, Rhona Patterson

Italian National Institute of Health

ISS (I)

Dr Gianfranco Brambilla, Dr Gabriele Vaccari

 

School of Agriculture and Forestry Engineering of the University of Cordoba

UCO (Sp)

Prof Dr Ana Garrido Varo, Ir Ma Dolores Perez Marin

Prof Dr Augusto Gomez Cabrera, Prof Dr José Emilio Guerrero Ginel

Laboratory of the autonomous government of Catalonia

LAGC (Sp)

Dr Jaume Bosch, Dr Silvia Termes