CEI Moncloa as the Origin of this Interdisciplinary Research

The main outcome of the think tank was the realisation that there was a need (both in biomedicine and the Agrofood industry) for low-cost real-time non-destructive methods for detecting peanut traces in powdered food ingredients such as flour, milk and cocoa. According to IUPAC^[1], "trace" means any element having an average concentration of less than about 100 parts per million atoms (ppm) or less than 100 ug g-1. One of the relevant issues for the development of a new procedure is the need of a reliable reference procedure which should be at least one order of magnitude more sensitive than the proposed one. In this research the reference procedure was developed (prior to CEI_Moncloa) by the UCM research group (TRADETBIO) and consisted of both PCR and real-time PCR [1,2] for the quantification of peanut traces with 0.1 ppm resolution.

On the other hand, the spatial resolution of the Hyper Spectral images (HSI) images proposed by UPM (LPF_Tagralia) was identified as the cornerstone for the detection of peanut traces, therefore at the very initial steps the optimal configuration of HIS was set as to characterize particles of 70um size.

Selection of Wavelength Range

VIS and NIR spectroscopy and HIS have been evaluated in consecutive trials [3-9] using IRMM^[i]-481e originating from Jumbo Runners, USA to guarantee the highest variability among peanut origins and treatments. The main conclusion being that VIS is sensitive to peanut variability, and does not provide sufficient specificity regarding milk and flour, based on results validated with commercial peanut samples (Figure 2). On the other hand, NIR spectroscopic imaging provides characteristic bands for peanut detection: 950 nm arising from O-H bond, 1212 nm arising from 2nd overtone of the C-H stretch of CH2group and 1450 nm arising from the 1st overtone of the O-H stretch, while HIS measurements at 70 µm resolution has shown the feasibility to detect up to 100 ppm (0.01 %) of peanut traces in any of the proposed powder foods (Figure 3).

SPECIFICITY ANALYSIS

Detecting peanut traces is relevant when it is sufficiently specific. Therefore, it is mandatory to avoid both false positives and false negatives. To this end, 49 commercial samples of cereals, legumes, oilseeds and nutswere obtained from local market of Madrid, Spain. All samples were ground with precautions to avoid cross-contamination. After grinding, all powdered food sampleswere sifted by passing through a sieve of 212 um, since diffuse reflectanceproperties are dependent on the size of particles, non-uniformity in size can cause scatter effects in the spectra. In this case the wavelength range proposed included all NIR data, except for the bands with highest sensitivity to peanuts. The result (see procedure in Figure 5a), attained for calibration and external validation [10,11] showed that the greatest challenge was in isolating peanuts from pine nuts. Figure 4b provides an indication or the most relevant wavelength bands for specificity. 

Calibration Transfer, Transcontinental Approach

Finally, it was necessary to determine whether our results were reproducible irrespective of the instrumental platform, a hypothesis which was dismissed after initial trials at the transcontinental level [12]. Therefore the development of calibration transfer techniques is required, and this step is still ongoing. Figure 5b summarizes some pros and cons of NIR spectroscopic techniques for the specific identification and quantification of peanut traces in powdered foods.

1. Maria Lopez-Calleja I, de la Cruz S, Gonzalez I, Garcia T, Martin R (2014) Survey of undeclared allergenic pistachio (Pistaciavera) in commercial foods by hydrolysis probe real-time PCR. Food Control 39: 49-55.
2. Maria Lopez-Calleja I, de la Cruz S, Pegels N, Gonzalez I, Garcia T, et al. (2013) High resolution TaqMan real-time PCR approach to detect hazelnut DNA encoding for ITS rDNA in foods. Food Chem 141: 1872-1880.
3. Mishra P, Barreiro Elorza P, Roger JM, Diezma Iglesias B, HerreroLangreo A, et al. (2014) Detection of peanut traces in wheat flour through NIR hyperspectral imaging spectroscopy. Hyperspectral imaging conference. Ricoh arena, Coventry, United Kingdom.
4. Mishra P, Roger J-M, Barreiro Elorza P, Diezma Iglesias B, HerreroLangreo A, LleóGarcía L, et al. (2014) Application of Improved Direct Calibration for Hyperspectral image processing: Detecting peanut traces in wheat flour. Food Analysis Congress, Barcelona, Spain.
5. Mishra P, Herrero-Langreo A, Barreiro P, Roger JM, Diezma B, et al. (2015) Detection and quantification of peanut traces in wheat flour by near infrared hyperspectral imaging spectroscopy using principal-component analysis. J Infrared Spectroscopy 23: 15-22.
6. [Mishra P, HerreroLangreo A, Barreiro Elorza P, Roger J-M, Diezma Iglesias B, et al. (2015) Hyperspectral to multispectral imaging for detection of tree nuts and peanut traces in wheat flour. J Spectral Imaging 4: 1-9.
7. Mohamad SNH, Ghosh S, Mishra P, Barreiro Elorza P, Diezma Iglesias B. (2015) Star dust memories, or how to detect allergen traces in powder foods. VII Congreso de EstudiantesUniversitarios de Ciencia, Tecnología e IngenieríaAgronómica. Madrid, Spain.
8. Ghosh S, Domínguez TRC, Diezma B, Lleó L, Barreiro P, et al. (2015) VIS/NIR spectral signature for the identification of peanut contamination of powder foods. AgricEngInt J 17: 310-329.
9. Cuadrado Dominguez TR, Ghosh S, Barreiro P, Lleo Garcia L, Diezma B, et al. (2014) VIS-NIR spectroscopy for the detection of peanuts traces in powder food. In: Tellez FA, Rodriguez AM, Sancho IM, Robinson MV, RuizAltisent M, Ballesteros FR, Hernando ECC, eds. VII CongresoIbérico de Agroingenieria y CienciasHorticolas: Innovar y Producirpara el Futuro Innovating and Producing for the Future. Madrid, Spain.
10. Mishra P, Cordella CBY, Rutledge DN, Barreiro P, Roger JM, et al. (2016) Application of independent components analysis with the JADE algorithm and NIR hyperspectral imaging for revealing food adulteration. J Food Eng 168: 7-15.
11. Ghosh S, Mishra P, Mohamad SNH, Martin de Santos R, Diezma Iglesias B, et al. (2016) Discrimination of peanuts from bulk cereals and nuts by near infrared reflectance spectroscopy. BiosystemsEng 151: 178-186.
12. Cuadrado Dominguez TR, Barreiro Elorza P, LleóGarcía L, Ruiz F, et al. (2016) Prospective of NIR Calibration transfer need for detecting peanut traces in powder foods: a transcontinental approach. 4th CIGR International Conference of Agricultural Engineering (CIGR-AgEng2016). Aahrus University, Denmark.