NIR spectroscopy and chemometrics for the typification of Spanish wine vinegars with a protected designation of origin

Depósito de Investigación de la Universidad de Sevilla https://idus.us.es/ This is an Accepted Manuscript of an article published by ELSEVIER in FOOD CONTROL, Vol. 89, on 2018, available at: https://doi.org/10.1016/j.foodcont.2018.01.031 Copyright 2018 Elsevier. En idUS Licencia Creative Commons CC BY-NC-ND Accepted Manuscript NIR spectroscopy and chemometrics for the typification of Spanish wine vinegars with a protected designation of origin Rocío Ríos-Reina, Diego Luis García-González, Raquel María Callejón, José Manuel Amigo PII: S0956-7135(18)30043-4 DOI: 10.1016/j.foodcont.2018.01.031 Reference: JFCO 5962 To appear in: Food Control Received Date: 12 December 2017 Revised Date: 29 January 2018 Accepted Date: 30 January 2018 Please cite this article as: Ríos-Reina Rocí., García-González D.L., Callejón Raquel.Marí. & Amigo José.Manuel., NIR spectroscopy and chemometrics for the typification of Spanish wine vinegars with a protected designation of origin, Food Control (2018), doi: 10.1016/j.foodcont.2018.01.031. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT 1 NIR SPECTROSCOPY AND CHEMOMETRICS FOR THE TYPIFICATION OF SPANISH WINE VINEGARS WITH A PROTECTED DESIGNATION OF ORIGIN Rocío Ríos-Reina a , Diego Luis García-González b , Raquel María Callejón a , José Manuel Amigo c,d* a Departamento de Nutrición y Bromatología, Toxicología y Medicina Legal. Facultad de Farmacia, Universidad de Sevilla, C/P. García Gonzalez no. 2, E-41012 Sevilla, Spain. b Instituto de la Grasa (CSIC), Campus University Pablo de Olavide - Building 46, Ctra. de Utrera, km. 1 E– 41013, Sevilla, Spain. c Department of Food Sciences, Spectroscopy and Chemometrics, Faculty of Sciences, Univ. Copenhagen, Rolighedsvej 30, DK-1958 Frederiksberg C, Denmark. d Department of Fundamental Chemistry, Federal University of Pernambuco, Av. Prof. Moraes Rego, 1235 – Cidade Universitária, Recife, Brazil; * e-mail corresponding author: [email protected]k 1 MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT 2 Abstract 2 High-quality wine vinegars protected by the indication “Protected Designation of Origin” 3 (PDO) need efficient tools to protect their brands and prevent adulteration and unfair 4 competition. In this sense, Near-Infrared spectroscopy (NIRs) combined with chemometrics has 5 demonstrated its usefulness in food authentication. This work assessed NIRs and Chemometrics 6 as a rapid and non-destructive methodology for this purpose. In this study, 83 high-quality wine 7 vinegars of the Spanish PDOs “Vinagre de Jerez”, “Vinagre de Condado de Huelva” and 8 “Vinagre de Montilla-Moriles” of different categories, and 11 wine vinegars without PDO, were 9 analyzed in the range 12000-4000 cm -1 . Principal component analysis (PCA) was performed to 10 explore the spectra and Partial Least Squares-Discriminant Analysis (PLS-DA) was used to 11 build classification models. The high ability of prediction obtained (>90% correct classification) 12 demonstrated the usefulness of this methodology for authentication of PDO wine vinegars and 13 their categories. 14 15 Keywords: Wine vinegar; Protected Designation of Origin; Near-Infrared spectroscopy; 16 Principal component analysis; Partial Least Squares-Discriminant Analysis. 17 18 MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT 3 1. INTRODUCTION 19 Wine vinegar has become a highly appreciated food product in gastronomy and one of the 20 most consumed types of vinegar in Europe (Paneque, Morales, Burgos, Ponce, & Callejón, 21 2017). Some wine vinegars, traditionally linked to a specific geographical area, have their 22 specifications related to their chemical and sensory features controlled by European regulations 23 under a legislative system named “Protected Designation of Origin” (PDO) (Chinnici et al., 24 2009). Thus, as occurring with other food, such as extra virgin olive oil, wine vinegars with a 25 PDO are recognized as a food product with the highest quality. In this field, Spain is one of the 26 major producers of high-quality wine vinegars, producing three of the five types of PDO 27 vinegars in Europe (Council Regulation (EC) No 510/2006): “Vinagre de Jerez”, “Vinagre de 28 Condado de Huelva” and “Vinagre de Montilla-Moriles”. These vinegars are made under 29 traditional processed and from high quality wines protected by their corresponding PDO. 30 Furthermore, some of these PDO wine vinegars are subjected to a period of aging in wooden 31 butts causing chemical modifications in their composition (Morales, Tesfaye, García-Parrilla, 32 Casas, & Troncoso, 2002). According to the sweetness, time and system of aging (“criaderas 33 and solera” or “añada” systems), different categories are considered within each Spanish PDO 34 (Table 1) having singular and specific characteristics (Council Regulation (EC) No 510/2006). 35 Due to the demand of high-quality vinegars has significantly increased over the last years, 36 and in addition to food quality is directly related to commercial value, there are suspicious that 37 adulteration and unfair competition in the vinegar industry is being practiced (Consonni, 38 Cagliani, Rinaldini, & Incerti, 2008; Sáiz-Abajo, González-Sáiz, & Pizarro, 2004; Tesfaye, 39 Morales, García-Parrilla, & Troncoso, 2002b). For this reason, wineries and regulatory councils 40 are demanding effective analytical tools to allow rapid and inexpensive analysis to verify the 41 origin of the vinegars in order to protect their brands and to prevent from adulteration. Some of 42 the classical analytical methods suggested for assessing food quality and differentiating 43 geographical origins such as gas-chromatography-mass spectrometry (GC-MS) (Callejón, 44 Morales, Silva Ferreira, & Troncoso, 2008), atomic absorption spectrometry or high-45 MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT 4 performance liquid chromatography (HPLC) (Tesfaye, Morales, García-Parrilla, & Troncoso, 46 2002a), are based on the measurement of the chemical compounds presented in vinegar (e.g. 47 volatiles and phenolic compounds or metals). Although these methods provide high quality 48 information, they require sample pretreatment steps, and they are destructive, time-consuming 49 and expensive. For this reason, there is a growing interest in developing rapid, accurate, 50 inexpensive and non-destructive methodologies based on non-targeted techniques for 51 characterization and authentication of high-quality vinegars (Callejón et al., 2012; De la Haba, 52 Arias, Ramírez, López, & Sánchez, 2014; Fan et al., 2011). In this sense, vibrational 53 spectroscopic techniques, such as Near Infrared spectroscopy (NIRs) and Fourier Transform 54 mid infrared spectroscopy (FTIR) have demonstrated to meet these characteristics being 55 informative at molecular level and very useful for identification and verification of raw 56 materials and final products, producing a single spectral fingerprint of each matrix, and 57 moreover, enable the direct measurement of wine vinegar samples with minimum or no sample 58 preparation (Lohumi, Lee, Lee, & Cho, 2015). Thus, a previous study of the characterization of 59 Spanish PDO wine vinegars by FTIR spectroscopy equipped with an attenuated total reflectance 60 accessory (ATR) (Ríos-Reina, Callejón, Oliver-Pozo, Amigo, & García-González, 2017b) 61 demonstrated the usefulness of this technique in the control of the different categories described 62 in wine vinegar PDOs. This method was applied at first due to it provides a greater amount of 63 chemical information compared to NIR spectroscopy in terms of chemical assignment of 64 observances and allows the interpretation of the spectra without the need of complex 65 chemometrics. Nevertheless, although a direct identification of the compounds is difficult by 66 NIRs, this methodology is extremely useful for highlighting groups of compounds that have 67 more relevance, giving a fingerprint of each sample, as well as it is faster, easier to implement 68 and easy to use (Baeten & Dardenne, 2002; Karoui & De Baerdemaeker, 2007; Stuart, 2004). 69 For all these reasons, NIRs could be a good alternative to be applied and its performance in 70 high-quality wine vinegars authentication must be checked. 71 MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT 5 However, the fact that the differences between the NIR spectra of different compounds are 72 usually very subtle, and the spectral occurrences in the NIR region are commonly dominated by 73 overtones, combination absorption bands and normally possess broad overlapping, makes 74 necessary the use of chemometrics. Thus, the information provided by NIRs requires advanced 75 multivariate data analysis (such as principal component analysis ‘PCA’, classification methods 76 as partial least squares-discriminant analysis ‘PLS-DA’, soft independent modelling by class 77 analogy ‘SIMCA’, etc.) to allow an efficient treating and interpreting of the signals, as well as 78 to perform a discrimination, classification and authentication of samples. In this context, in the 79 last few years, the use of NIRs in combination with multivariate chemometric analysis has been 80 widely reviewed for many different approaches such as authentication, detecting adulteration or 81 differentiating geographical origins of food products (Cozzolino, 2014; Grassi, Amigo, 82 Lyndgaard, Foschino, & Casiraghi, 2014; Pillonel et al., 2003; Alamprese, Amigo, Casiraghi, & 83 Engelsen, 2016; Liu et al., 2008). However, with regard to the classification of vinegars by 84 NIRs, there are only a few papers related to wine vinegars and high-quality wine vinegars, 85 pointing out the utility of this spectroscopic technique in relation to other food commodities 86 (Casale, Sáiz Abajo, González Sáiz, Pizarro, & Forina, 2006; Fan et al., 2011; Zhao, Zhang, 87 Zhao, Zhang, & Liu, 2011). Furthermore, despite the advantages of NIRs and FTIR vibrational 88 spectroscopic techniques are well known nowadays, the implementation of one of these 89 techniques in the characterization and classification of these high-quality wine vinegars still 90 requires to be further studied, comparing their suitability in the analysis of this food matrix. For 91 these reasons, the aim of the study is to investigate the potential and suitability of NIRs in 92 conjunction with multivariate classification tools as a rapid, inexpensive and non-destructive 93 methodology for the characterization and authentication of the three Spanish wine vinegar 94 PDOs (“Vinagre de Jerez”, “Vinagre de Condado de Huelva” and “Vinagre de Montilla-95 Moriles”), assessing its ability to classify PDO wine vinegars according to their category and to 96 discriminate them from commercial wine vinegars without PDO. 97 98 MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT 6 2. MATERIALS AND METHODS 99 2.1 Samples 100 2.1.1 PDO wine vinegars 101 Eighty-three wine vinegar samples belonging to the three Spanish PDOs were analyzed 102 in this study: 41 samples from “Vinagre de Jerez”, 29 from “Vinagre de Condado de Huelva”, 103 and 13 from “Vinagre de Montilla-Moriles”. These samples were provided by the Regulatory 104 Councils of each Spanish PDO, which assessed and certified the authenticity of the wine 105 vinegars. The less number of samples collected for the PDO “Vinagre de Montilla-Moriles”, as 106 well as the lack of “Gran Reserva” samples, is explained by the fact that it has been recently 107 registered with the indication of PDO (registered in 2015). Furthermore, a different number of 108 samples within each PDO were collected for the established categories (aged and sweet) due to 109 the rate of production of each category during last years (2014-2015). More information about 110 samples included in the study is shown in Table 1. 111 2.1.2 Commercial wine vinegar samples without PDO 112 A total amount of 11 wine vinegars from different regions were purchased in local markets 113 and wineries and named in the study as “Commercial samples without PDO” (V): 7 samples 114 produced in northern Spain; 1 wine vinegar from the same region as “Vinagre de Montilla-115 Moriles” PDO but without the PDO indication; and 3 samples without specification of the 116 geographical origin. The number of samples was inevitably limited by their production and 117 availability. Therefore, the work was developed under a feasibility point of view and the 118 production and occurrence in the market was take into account for the construction of the 119 models. 120 2.2 NIR measurements 121 NIR spectra were collected in absorption mode using an ABB Bomen IR spectrometer (Q-122 interline, X, Denmark), equipped with a 1 mm path length cuvette. Spectral data were collected 123 in the range of 12000–4000 cm -1 , with a resolution of 8 cm -1 and 64 scans for both backgrounds 124 and samples. Wine vinegar samples were directly analyzed without sample pre-treatment by 125 MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT 7 pipetting them into 1 mL shell vial, 40x80 mm transparent (Skandinaviska Genetec AB, Lund, 126 Sweden) before measurement. The spectrometer was interfaced to a computer with 127 GRAMS/AI™ Spectroscopy Software (Thermo Fisher Scientific software) for spectral 128 acquisition and exportation. The spectrum of each sample was obtained in triplicate in a random 129 sequence at room temperature (21–23 °C). 130 2.3 Data processing and multivariate analysis 131 Data analysis was performed by using PLS_Toolbox 7.9.5 (Eigenvector Research Inc., 132 Wenatchee, WA) working under MATLAB v.8.5.0 environment (The Mathworks Inc., Natick, 133 MA). Different preprocessing methods were studied prior to multivariate data analysis. The best 134 pre-processing method was smoothing (SMT) 7 point and second order filtering operation, to 135 reduce random noise and standard normal variate (SNV) method (Barnes, Dhanoa, & Lister, 136 1989) to correct for baseline variations due to the different scattering of the samples. Moreover, 137 mean centering (MC) was performed on the spectra. Two segments of the spectrum were 138 removed from the whole wavenumber range of the spectra: the first one because of the low 139 value of the signal/noise and the second one because of the strong combination band of O-H 140 from water (4000-5430 cm -1 and 7200-6400 cm -1 , respectively). The corrected NIR spectra 141 before and after preprocessing are shown on Fig. I (Supplementary Material). 142 Before classification models, an exploratory analysis of the data is advisable to be 143 performed to detect outliers, recognize patterns in samples distribution and relationships 144 between variables and classes. For this purpose, PCA was carried out prior to any classification 145 approach. After the PCA models, full cross validation (leave-one-out) was used as validation 146 method for the PLS-DA models. Several PLS-DA models were built with different 147 classification purposes: the first one, classifying the different commercialized categories (aged 148 and sweet) within the same PDO, and the second purpose was to differentiate wine vinegars 149 with PDO from those without PDO certification. The models were tested using a data set that 150 was not used in the process of calibration model building. The samples belonging to each 151 dataset were randomly selected by the Kennard-Stone algorithm (Kennard & Stone, 1969) and 152 MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT 14 a first step, different PCA models were built (Fig. II Supplementary Material). The score plots 298 showed a clear difference between the PDO wine vinegars and the group of vinegars without 299 PDO. Only the visual differentiation between some “Pedro Ximenez” samples belonging to 300 “Vinagre de Montilla-Moriles” PDO (MPX) and one “Pedro Ximenez” wine vinegar without 301 PDO (VPX) was not perfectly clear (Fig. II C.1 Supplementary Material). However, a “Pedro 302 Ximenez” sample without PDO that was produced in the same geographical area as “Vinagre de 303 Montilla-Moriles” PDO (VMPX) was placed in the scores plot extremely separated from 304 “Pedro Ximenez” vinegars within the PDO (MPX). These results reaffirms the unique quality 305 and characteristics of wine vinegars produced under the specifications of a PDO due to major 306 controls and their traditional method of production that provided high-quality conditions to 307 vinegars since a very long period of time is required. No use of NIRs technology on the 308 differentiation of PDO wine vinegars from vinegars without the PDO indication has been 309 reported to date. 310 Regarding loadings plot (Fig. II A.2, B.2, C.2 Supplementary Material), the first two 311 PCs, which explained between 94% and 99% of total variance in the three PCA models, pointed 312 out that the spectral regions mainly responsible of the differentiation were again those between 313 5000-6500 cm -1 together with the region between ~10000 cm -1 and ~12000 cm -1 that had an 314 important relevance in this particular case. 315 3.2. PLS-DA classification models 316 After a preliminary exploratory analysis of spectra by PCA, PLS-DA model was developed 317 for a classification purpose. The first PLS-DA models were developed to classify samples 318 between their established PDO categories (henceforth “category classification”). The second 319 PLS-DA models were developed to confirm the ability of NIRs to authenticate and differentiate 320 PDO wine vinegars from wine vinegars without the PDO designation (henceforth “PDO/origin 321 classification”). Categories with lower number of samples were not included in the models. 322 Moreover, “Añada” category was grouped with “Reserva” category in “Vinagre de Condado 323 de Huelva” PDO, as the aging time regulated in each one was similar (more than two-three 324 MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT 15 years of aging), differing only on the aging system used. The models were tested by dividing the 325 total number of samples in two sets (training and test sets). Further information about the 326 number of samples used for modeling and predicting are shown in Table I Supplementary 327 Material. 328 3.2.1. PLS-DA models for distinguishing the three aged categories within each PDO 329 (category classification) 330 The statistical parameters obtained by PLS-DA in the different models are shown in Table 331 2a. High sensitivity and specificity values (%) were obtained in the PLS-DA models for each 332 category. The 87-100% of the samples were correctly classified, demonstrating that all the 333 categories within each PDO could be successfully separated from the rest of classes. These 334 results confirmed and improved those obtained in a previous study of the authors (Ríos-Reina et 335 al., 2017a), in which these vinegars were analyzed by multidimensional fluorescence 336 spectroscopy coupled with different classification tools, resulting in the need of using a non-337 linear classification tool (support vector machines) to obtain good classification results. In the 338 case of NIRs, a linear classification approach is enough for obtaining good results. However, the 339 results also showed that the categories that showed lowest classification rates were those that 340 would be in the boundaries between categories according to their aging period (“Solera” for 341 “Vinagre de Condado de Huelva” PDO and “Reserva” in the other two PDOs). These results 342 were acceptable considering the high variability of these samples due to the factor that they 343 were aged over a wide range of time that is reflected over their complex chemical composition 344 (García-Parrilla et al., 1999). These intermediate categories were expected to be 345 spectroscopically and chemically similar to the vinegars of the immediately previous or 346 following category. Furthermore, the highly variability in the intermediate categories was also 347 observed by Callejón et al., (2012), whose study revealed that the lowest classification rates 348 were obtained for the intermediate aged category “Reserva”, with an aging time between 349 “Crianza” and “Gran Reserva” categories. 350 MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT 16 3.2.2. PLS-DA classification of PDO wine vinegars and wine vinegars without PDO 351 (PDO/origin classification) 352 After the exploratory PCA analysis, a PLS-DA was applied to confirm the ability of 353 NIRs to authenticate and differentiate PDO wine vinegars from those without the PDO 354 indication. PLS-DA results are shown in Table 2b. The low classification errors of prediction 355 obtained in the models demonstrated that a good separation of PDO wine vinegar samples from 356 those without the PDO certification could be performed with the proposed methodology. As the 357 number of samples between the two groups (with and without PDO) was not properly balanced 358 for building robust models, several PLS-DA models were developed and tested with the same 359 number of samples per group. Samples from the PDO group (11 different samples each time) 360 were randomly selected and included in the models together with samples without PDO. The 361 results obtained matched with those shown in Table 2b. These results highlighted the unique 362 characteristics conferred by the high quality of raw wines used (each belonging to the 363 corresponding PDO), the traditional system of production and aging of the Spanish PDO wine 364 vinegars (“criaderas and solera” or “añada” systems), and the standardize procedure of 365 production. All of these characteristics, together with the routine controls by the regulatory 366 councils, allowed a rapid classification and differentiation from the rest of wine vinegars 367 without a PDO indication. Although other researches showed the utility of NIRs in the 368 differentiation of vinegars with different raw materials (Sáiz-Abajo et al., 2004), or even 369 between different wine vinegar manufacturing methods (De la Haba et al., 2014), no references 370 have been found showing the differentiation and classification of Spanish PDO wine vinegars 371 from vinegars without a PDO by using only NIRs. 372 3.2.3. Comparison of classification results obtained between NIRs and FTIR analysis 373 In order to explore the potential and advantages of using NIRs and ATR-FTIR 374 spectroscopy and their suitability for PDO wine vinegar classification, the performance of the 375 two techniques was compared for a wine vinegar classification purpose (classification of 376 categories within a PDO). For this purpose, PLS-DA results obtained by both spectroscopic 377 MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT 17 techniques were examined by comparing the percentage of correct predictions (Table 3). For 378 this comparison, PLS-DA classification models were built with ATR-FTIR data obtained in a 379 previous research carried out by the authors of this study (Ríos-Reina, Callejón, Oliver-Pozo, 380 Amigo, & García-González, 2017b). In this model, the range between 1500 and 900 cm -1 was 381 included in the PLS-DA, due to the fact that, as it was previously reported, it showed the main 382 spectral bands assigned to complex interacting vibrations related to the unique fingerprint of 383 each vinegar. 384 NIR classification models showed percentages of correct predictions in the range 86.7-385 100% in most of the categories while in the case of ATR-FTIR the percentage of correct 386 prediction were 58.4-100% (Table 3). In most of the cases, the classification rates were higher 387 in NIRs compared to ATR-FTIR. However, it is important to consider the advantages of both 388 techniques. Thus, ATR-FTIR spectroscopy has the advantages of being able to determine 389 absorption bands with clear chemical assignments, which facilitates the interpretation of the 390 spectra. Although NIR spectra are more difficult to interpret and the calibration procedures were 391 more complicated, it also shows an easy and robust analysis and yielded satisfactory 392 classification results for PDO wine vinegars. Depending on the classification purpose 393 (categories and PDO vs non-PDO) and the needs for interpreting the spectra, one of both 394 techniques could be proposed to be applied in quality control of vinegars, or even the 395 combination of both of them. 396 397 MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT 18 4. Conclusions 398 In this work, the combination of NIR with chemometrics has demonstrated to be useful 399 for a rapid characterization and classification of the Spanish PDO wine vinegars and for 400 controlling the authenticity of their commercialized categories (aged and sweet). A simple 401 exploration of the NIR data by a PCA pointed out some that aging and the protection under a 402 PDO had an effect in the spectra, showing similarities between the spectra of the aged 403 categories of the three Spanish PDOs. The absorption bands most involved in aging changes, 404 and also related to sweet category, were those from ~5200 to ~6500 cm -1 , associated to the 405 presence of water and aromatic and phenolic compounds that have shown changes during aging. 406 Furthermore, the sweet category “Pedro Ximenez” showed some characteristic bands at the 407 same region (~5600 cm-1) mainly associated to sugars, due to their special characteristic of 408 production. The unique characteristics of the Spanish PDO wine vinegars, which directly affect 409 to the NIR spectra, allowed a satisfactory classification according to the category (aged and 410 sweet categories within each PDO) and PDOs versus non-PDO differentiation (PDO wine 411 vinegars from vinegars without this quality indication) by the development of PLS-DA 412 classification models with the NIR spectrum of samples. 413 The advantages of this methodology would allow implementing it as an alternative tool 414 for fingerprinting wine vinegar samples on a large scale, this analytical tool being cost-effective 415 and rapid. Further research will be carried out to test this technique at industrial scale with a 416 higher number of samples to evaluate the efficiency in real authentication problems. 417 5. Acknowledgements 418 The authors would like to thank the Spanish Regulatory Councils of the wine vinegars 419 PDOs for their invaluable help with the acquisition of the samples for the study. Authors also 420 thank University of Copenhagen, specifically to Professor Franciscus Winfried J van der Berg 421 and the department of food science for providing the equipment and their knowledge in the 422 methodology. This work was supported by “Consejeria de Economía, Innovación y Ciencia” of 423 MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT 19 “Junta de Andalucia” [P12-AGR-1601], and “Ministerio de Educación Cultura y Deporte” 424 (MECD, Spanish Government) for the FPU pre-doctoral fellowship [grant number 425 FPU2014/01247]. 426 427 6. References 428 Alamprese, C., Amigo, J. M., Casiraghi, E., & Engelsen, S. B. (2016). Identification and 429 quantification of turkey meat adulteration in fresh, frozen-thawed and cooked minced beef 430 by FT-NIR spectroscopy and chemometrics. Meat Science, 121, 175–181. 431 http://doi.org/10.1016/j.meatsci.2016.06.018 432 Baeten, V., & Dardenne, P. (2002). Spectroscopy: Developments in instrumentation and 433 analysis. Grasas Y Aceites, 53(1), 45–63. http://doi.org/10.3989/gya.2002.v53.i1.289 434 Barnes, R. J., Dhanoa, M. S., & Lister, S. J. (1989). Standard Normal Variate Transformation 435 and De-trending of Near-Infrared Diffuse Reflectance Spectra. Applied Spectroscopy, 43, 436 772–777. 437 Callejón, R. M., Amigo, J. M., Pairo, E., Garmón, S., Ocaña, J. A., & Morales, M. L. (2012). 438 Classification of Sherry vinegars by combining multidimensional fluorescence, parafac 439 and different classification approaches. Talanta, 88, 456–462. 440 http://doi.org/10.1016/j.talanta.2011.11.014 441 Callejón, R. M., Morales, M. L., Silva Ferreira, A. C., & Troncoso, A. M. (2008). Defining the 442 typical aroma of Sherry vinegar: Sensory and chemical approach. Journal of Agricultural 443 and Food Chemistry, 56(17), 8086–8095. http://doi.org/10.1021/jf800903n 444 Callejón, R. M., Torija, M. J., Mas, A., Morales, M. L., & Troncoso, A. M. (2010). Changes of 445 volatile compounds in wine vinegars during their elaboration in barrels made from 446 different woods. Food Chemistry, 120, 561–571. 447 Casale, M., Sáiz Abajo, M. J., González Sáiz, J. M., Pizarro, C., & Forina, M. (2006). Study of 448 MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT 20 the aging and oxidation processes of vinegar samples from different origins during storage 449 by near-infrared spectroscopy. Analytica Chimica Acta, 557(1–2), 360–366. 450 http://doi.org/10.1016/j.aca.2005.10.063 451 Chinnici, F., Guerrero, E. D., Sonni, F., Natali, N., Marín, R. N., & Riponi, C. (2009). Gas 452 chromatography-mass spectrometry (GC-MS) characterization of volatile compounds in 453 quality vinegars with protected European geographical indication. Journal of Agricultural 454 and Food Chemistry, 57(11), 4784–4792. http://doi.org/10.1021/jf804005w 455 Consonni, R., Cagliani, L. R., Rinaldini, S., & Incerti, A. (2008). Analytical method for 456 authentication of Traditional Balsamic Vinegar of Modena. Talanta, 75(3), 765–769. 457 http://doi.org/10.1016/j.talanta.2007.12.005 458 Council Regulation (EC) No 510/2006 of 20 March 2006 on the protection of geographical 459 indications and designations of origin for agricultural products and foodstuffs. http://eur-460 lex.europa.eu/legal-content/EN/ALL/?uri=CELEX%3A32006R0510. Accessed 15 March 461 2016. 462 Cozzolino, D. (2014). An overview of the use of infrared spectroscopy and chemometrics in 463 authenticity and traceability of cereals. Food Research International, 60, 262–265. 464 http://doi.org/10.1016/j.foodres.2013.08.034 465 Cozzolino, D., Smyth, H. E., & Gishen, M. (2003). Feasibility Study on the Use of Visible and 466 Near-Infrared Spectroscopy Together with Chemometrics to Discriminate between 467 Commercial White Wines of Different Varietal Origins. Journal of Agricultural and Food 468 Chemistry, 51(26), 7703–7708. http://doi.org/10.1021/jf034959s 469 De la Haba, M. J., Arias, M., Ramírez, P., López, M. I., & Sánchez, M. T. (2014). 470 Characterizing and authenticating montilla-moriles PDO vinegars using near infrared 471 reflectance spectroscopy (nirs) technology. Sensors (Switzerland), 14(2), 3528–3542. 472 http://doi.org/10.3390/s140203528 473 MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT 21 Fan, W., Li, H., Shan, Y., Lv, H., Zhang, H., & Liang, Y. (2011). Classification of vinegar 474 samples based on near infrared spectroscopy combined with wavelength selection. 475 Analytical Methods, 3(8), 1872. http://doi.org/10.1039/c1ay05101f 476 García-Parrilla, M. C., Heredia, F. J., & Troncoso, A. M. (1999). Sherry wine vinegars: 477 Phenolic composition changes during aging. Food Research International, 32(6), 433–478 440. http://doi.org/10.1016/S0963-9969(99)00105-2 479 Giangiacomo, R., & Dull, G. G. (1986). Near Infrared Spectrophotometric Determination of 480 Individual Sugars in Aqueous Mixtures. Journal of Food Science, 51(3), 679–683. 481 http://doi.org/10.1111/j.1365-2621.1986.tb13910.x 482 Grassi, S., Amigo, J. M., Lyndgaard, C. B., Foschino, R., & Casiraghi, E. (2014). Beer 483 fermentation: Monitoring of process parameters by FT-NIR and multivariate data analysis. 484 Food Chemistry, 155, 279–286. http://doi.org/10.1016/j.foodchem.2014.01.060 485 Karoui, R., & De Baerdemaeker, J. (2007). A review of the analytical methods coupled with 486 chemometric tools for the determination of the quality and identity of dairy products. Food 487 Chemistry, 102(3), 621–640. http://doi.org/10.1016/j.foodchem.2006.05.042 488 Kennard, R.W. and Stone, L.A. (1969) Computer-aided Design of Experiments. Technometrics, 489 11, 137-148. https://doi.org/10.1080/00401706.1969.10490666 490 Liu, L., Cozzolino, D., Cynkar, W. U., Dambergs, R. G., Janik, L., O’Neill, B. K., Colby, C. B., 491 Gishen, M. (2008). Preliminary study on the application of visible-near infrared 492 spectroscopy and chemometrics to classify Riesling wines from different countries. Food 493 Chemistry, 106(2), 781–786. http://doi.org/10.1016/j.foodchem.2007.06.015 494 Lohumi, S., Lee, S., Lee, H., & Cho, B. K. (2015). A review of vibrational spectroscopic 495 techniques for the detection of food authenticity and adulteration. Trends in Food Science 496 and Technology, 46(1), 85–98. http://doi.org/10.1016/j.tifs.2015.08.003 497 Morales, M. L., Tesfaye, W., García-Parrilla, M. C., Casas, J. A., & Troncoso, A. M. (2002). 498 MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT 22 Evolution of the aroma profile of sherry wine vinegars during an experimental aging in 499 wood. Journal of Agricultural and Food Chemistry, 50(11), 3173–8. 500 http://doi.org/10.1021/jf011313w 501 Osborne, B., Fearn, T., & Hindle, P. (1993). Practical NIR spectroscopy with applications in 502 food and beverage analysis. (A.-W. Longman, Ed.) ((2nd ed. )). Ltd: Harlow UK. 503 Paneque, P., Morales, M. L., Burgos, P., Ponce, L., & Callejón, R. M. (2017). Elemental 504 characterisation of Andalusian wine vinegars with protected designation of origin by ICP-505 OES and chemometric approach. Food Control, 75, 203–210. 506 http://doi.org/http://dx.doi.org/10.1016/j.foodcont.2016.12.006 507 Pillonel, L., Luginbühl, W., Picque, D., Schaller, E., Tabacchi, R., & Bosset, J. O. (2003). 508 Analytical methods for the determination of the geographic origin of Emmental cheese: 509 Midand near-infrared spectroscopy. European Food Research and Technology, 216(2), 510 174–178. http://doi.org/10.1007/s00217-002-0628-5 511 Ríos-Reina, R., Callejón, R. M., Oliver-Pozo, C., Amigo, J. M., & García-González, D. L. 512 (2017b). ATR-FTIR as a potential tool for controlling high quality vinegar categories. 513 Food Control, 78, 230–237. http://doi.org/10.1016/j.foodcont.2017.02.065 514 Ríos-Reina, R., Elcoroaristizabal, S., Ocaña-González, J. A., García-González, D. L., Amigo, J. 515 M., & Callejón, R. M. (2017a). Characterization and authentication of Spanish PDO wine 516 vinegars using multidimensional fluorescence and chemometrics. Food Chemistry, 230, 517 108–116. http://doi.org/10.1016/j.foodchem.2017.02.118 518 Sáiz-Abajo, M. J., González-Sáiz, J. M., & Pizarro, C. (2004). Classification of wine and 519 alcohol vinegar samples based on near-infrared spectroscopy. Feasibility study on the 520 detection of adulterated vinegar samples. Journal of Agricultural and Food Chemistry, 521 52(25), 7711–7719. http://doi.org/10.1021/jf049098h 522 Stuart, B. H. (2004). Infrared Spectroscopy: Fundamentals and Applications. Methods (Vol. 8). 523 MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT 23 John Wiley & Sons, Ltd. http://doi.org/10.1002/0470011149 524 Tesfaye, W., Morales, M. L., Benítez, B., García-Parrilla, M. C., & Troncoso, a. M. (2004). 525 Evolution of wine vinegar composition during accelerated aging with oak chips. Analytica 526 Chimica Acta, 513(1), 239–245. http://doi.org/10.1016/j.aca.2003.11.079 527 Tesfaye, W., Morales, M. L., García-Parrilla, M. C., & Troncoso, A. M. (2002a). Evolution of 528 phenolic compounds during an experimental aging in wood of Sherry vinegar. Journal of 529 Agricultural and Food Chemistry, 50(24), 7053–7061. http://doi.org/10.1021/jf020602x 530 Tesfaye, W., Morales, M. L., García-Parrilla, M. C., & Troncoso, A. M. (2002b). Wine vinegar: 531 Technology, authenticity and quality evaluation. Trends in Food Science and Technology, 532 13(1), 12–21. http://doi.org/10.1016/S0924-2244(02)00023-7 533 Yu, H., Fu, X., Xie, L., Ying, Y., & Zhou, Y. (2007). Discrimination between Chinese rice 534 wines of different geographical origins by NIRS and AAS. European Food Research and 535 Technology, 225(3–4), 313–320. http://doi.org/10.1007/s00217-006-0416-8 536 Zhao, Y., Zhang, S., Zhao, H., Zhang, H., & Liu, Z. (2011). Fast discrimination of mature 537 vinegar varieties with Visible-NIR spectroscopy. IFIP Advances in Information and 538 Communication Technology, 344 AICT(PART 1), 721–728. http://doi.org/10.1007/978-3-539 642-18333-1_87. 540 541 MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT Highlights • Near-Infrared spectroscopy was studied for characterizing PDO wine vinegars • PLS-DA models were able to differentiate the Spanish PDOs and wine vinegar categories • Wine vinegars without PDO were used to test the models • NIR provided better classification results than ATR-FTIR analysis • These techniques, or their combination, could be useful in vinegar quality control