CLASSIFICATION OF CHINESE RICE WINE WITH DIFFERENT MARKED AGES BASED ON NEAR INFRARED SPECTROSCOPY ABSTRACT

Transcription

1 CLASSIFICATION OF CHINESE RICE WINE WITH DIFFERENT MARKED AGES BASED ON NEAR INFRARED SPECTROSCOPY HAIYAN YU, YIBIN YING 1, XIAPING FU and HUISHAN LU College of Biosystems Engineering and Food Science Zhejiang University Hangzhou, China ABSTRACT The feasibility of near infrared (NIR) spectroscopy for discrimination between Chinese rice wine with different marked ages (1, 3 and 5 years) was presented in this research. NIR spectra were collected in transmission mode in the wavelength range of nm. Discriminant models were developed based on discriminant analysis (DA) together with raw and second derivative spectra. The calibration result for raw spectra was better than that for second derivative spectra. The percentage of samples correctly classified for raw and second derivative spectra was 100 and 78.4%, respectively. In validation analysis, the percentage of samples correctly classified was 94.4%. For 1-, 3- and 5-year-old sample groups, the percentage of samples correctly classified was 88.3, 100 and 100%, respectively. The results demonstrated that NIR spectroscopy could be used as a rapid and reliable method for classification of Chinese rice wine with different marked ages. INTRODUCTION Age and marked age of Chinese rice wine are definitely defined in the Official Methods of Analysis for Chinese rice wine (GB/T ; GB means national standard; is the code of the two official methods; and 2000 is the year when the method was revised). The age of Chinese rice wine is the fixed number of year when the finished wine product was stored in the pottery. Marked age is defined as the claimed age of Chinese rice wine on the commercial label, which is the age-weighted average of Chinese rice wine with different age. Regarding Chinese rice wine with a marked age of 3 years or higher, GB/T states that the content of basic liquor of the marked age should be more than 50%. 1 Corresponding author. TEL: ; FAX: ; ybying@ zju.edu.cn Journal of Food Quality 29 (2006) All Rights Reserved. 2006, The Author(s) Journal compilation 2006, Blackwell Publishing 339

2 340 H. YU ET AL. The common fraudulent practice in the commercialization of Chinese rice wine is to sell young wine under the denomination of aged wine. The adulterated products do not represent any safety risk for human health, but they constitute a fraud for consumers and are unfair practices before other Chinese rice wine producers. Wine identification and classification, mainly in terms of variety, geographical origin, vintage and wine age, has received increasing attention during the past 10 years (Kallithraka et al. 2001; Soufleros et al. 2003; Saenz-Lopez et al. 2004). Discrimination of wine authenticity was attempted by using different types of compositional data including volatile compounds, trace elements, amino acids and phenolic compounds (Garcia-Jares et al. 1995; Baxter et al. 1997; Mayen et al. 1997; Soufleros et al. 1998; Rebolo et al. 2000; Dugo et al. 2005). All of the methods require sophisticated and expensive analytical techniques such as high-performance liquid chromatography, gas chromatography, atomic absorption spectroscopy, capillary gas chromatography, mass spectrometry or capillary zone electrophoresis, and so on. Near infrared (NIR) spectroscopy has gained wide acceptance in the field of food engineering mainly because of its suitability for recording the spectra of solid and liquid samples at low cost, without any pretreatment, and in a nondestructive way. It is also an easy-to-use, reliable and versatile analytical technique for determining different compounds in wine, as well as in many other food products (Saiz-Abajo et al. 2004). The NIR region of the electromagnetic spectrum lies between the visible and middle infrared regions and spans the wavelength range between 780 and 2526 nm (American Society for Testing and Materials). It contains information concerning the relative proportions of C H, N H and O H bonds, which are the primary structural components of organic molecules (Osborne et al. 1993). NIR spectroscopy has been used to monitor ethanol in alcohol fermentation processes (Cavinato et al. 1990), predict methanol in wine-fortifying spirit (Dambergs et al. 2002), evaluate ethanol content of alcoholic beverages (Gishen and Dambergs 1998; Leornardo et al. 2003), determine phenolic compounds (malvidin-3-glucoside, pigmented polymers and tannins) in red wine fermentations (Cozzolino et al. 2004), screen 15 parameters (alcoholic degree, volumic mass, total acidity, ph, volatile acidity, glycerol, total polyphenol index, reducing sugars, lactic, malic, tartaric and gluconic acids, color, tonality, total sulfur dioxide and free sulfur dioxide) in different types of wines (red, rose and white wines) (Urbano-Cuadrado et al. 2004) and to determine alcoholic degree, ph value, total acid, amino acid nitrogen and [degree symbol]brix in Cinese rice wine (Yu et al. 2006). It also has been applied to discriminate between samples of Australian commercial white wines of different varietal origins (Chardonnay and Riesling) (Cozzolino et al. 2003), and to discriminate between wine and alcohol vinegars (Saiz-Abajo et al. 2004).

3 CLASSIFICATION OF CHINESE RICE WINE USING NIR SPECTROSCOPY 341 However, no reports were found in relation to the use of NIR spectroscopy for classification of Chinese rice wine with different marked ages. The aim of this work was to investigate the applicability of NIR spectroscopy for discrimination between Chinese rice wine with different marked ages (1, 3 and 5 years). MATERIALS AND METHODS Samples Thirty-six bottles of Chinese rice wine were obtained from Tapai brand Shaoxing rice wine brewery and 33 bottles were obtained from Kuaijishan brand Shaoxing rice wine brewery (Shaoxing rice wine is well known by the generic name Shaoxing, a sort of Denomination of Origin. It is made from high-quality glutinous rice and wheat, as well as water from Mirror Lake in the region of Shaoxing, Zhejiang province.). The samples were of different marked ages (1, 3 and 5 years), and were of 2004 and 2005 vintages, respectively. The samples were divided randomly into calibration and validation sets. Among them, 51 samples were used for the calibration set and 18 samples were used for the validation set. Details are listed in Table 1. Spectral Measurements Samples taken from freshly opened bottles of wine were scanned in transmission mode ( nm) using a commercial spectrometer (Nexus TABLE 1. DISTRIBUTION OF CHINESE RICE WINE SAMPLES IN THE CALIBRATION AND VALIDATION SETS Commercial label Marked age (years) Number of samples Vintage Calibration set Tapai , , 2005 Kuaijishan , , Validation set Tapai , , 2005 Kuaijishan , ,

4 342 H. YU ET AL. FT-NIR, Thermo Nicolet Corporation, Madison, WI), which was equipped with an interferometer, an InGaAs detector and a wide-band light source (50 W; Quartz Tungsten Halogen). The samples were scanned in a 1-mm optical path-length rectangular quartz cuvette with air as reference at room temperature. A reference scan was taken once in every three sample scans. NIR spectra were collected using OMNIC software (Thermo Nicolet Corporation) and stored as absorbance. Spectral range was from 800 to 2500 nm, mirror velocity was cm/s, and resolution was 16/cm -1 in this work. The spectrum of each sample was the average of 32 successive scans. Reference Analysis Reference analysis for alcoholic degree, total acid and Brix, three major components of Chinese rice wine, was in accordance with the Official Methods of Analysis for Chinese rice wine and Official Methods of Analysis for Shaoxing rice wine (GB/T and GB ). All analyses were done in duplicate. For alcoholic degree measurement, a furnace, condenser, alcoholometer (20C, 0.2C) and thermometer (50C, 0.1C) were used. The difference between two measurements was less than 0.2% (v/v). For total acid measurement, an analytical balance, electrictitration instrument (accuracy of ±0.02 ph; ZD-2, Shanghai Scientific Instrument Co. Ltd, Shanghai, China) and a magnetic stirrer (JB-1A, Shanghai Scientific Instrument Co. Ltd) were used. The titer difference between two measurements for total acid was less than 0.05 ml. The Brix (%) parameter shows the concentration percentage of soluble solid content in a sample (water solution). The soluble solid content is the total of all the solids dissolved in the water, including sugar, protein, acids, etc., and the measurement reading is the total sum of these (Saiz-Abajo et al. 2004). Basically, Brix (%) is calibrated to the number of grams of sucrose contained in 100 g of sucrose solution. In this work, Brix measurement was performed using a digital refractometer (PR-101, Atago Co. Ltd, Tokyo, Japan) with a Brix measurement range of 0.0~45.0%, accuracy of ±0.2% and automatic temperature compensation system of 5 40C. Chemometrics and Data Analysis The raw and second derivative spectra were used for calibration analysis. The second derivative was used to reduce baseline variation and enhance the spectral features. Chemometric analysis was performed using the commercial software package, TQ Analyst v6.2.1 (Thermo Nicolet Corporation). Discriminant analysis (DA) was used to classify Chinese rice wine samples with different

5 CLASSIFICATION OF CHINESE RICE WINE USING NIR SPECTROSCOPY 343 marked ages. It was used to determine the classes of known samples that were most similar to an unknown sample by computing the unknown s distance from each class center in Mahalanobis distance. Principal component analysis (PCA) was performed before DA models were developed. PCA was performed in order to reduce the number of variables showing colinearity. Thus, the samples were in a new reduction k-dimensional space (k n). From the k factor score, Mahalanobis distance, which indicated how different a sample spectrum was from the average spectrum of the sample set, was calculated. It was defined by the following equation: 1 MD = ( t t ) S ( t t ) i i k i where MD i is the Mahalanobis distance; t i is the score vector of ith sample; t is the mean score vector of the sample set; and S k is the scores covariance matrix. The DA models were then externally validated by using them to predict the marked age of samples in the validation set. T RESULTS AND DISCUSSION Reference Analysis Data The statistical data for alcoholic degree, total acid and Brix of the 1-, 3- and 5-year-old sample groups are listed in Table 2. As shown in Table 2, alcoholic degree decreases, while concentrations of total acid and Brix increase as the marked age increases. In order to compare statistically, F-test was performed on the reference data of the three sample groups. As listed in Table 2, the F values of the three parameters were all higher than F 0.05, which indicates that the values of alcoholic degree, total acid and Brix of the 1-, 3- TABLE 2. STATISTICAL DATA FOR ALCOHOLIC DEGREE, TOTAL ACID AND BRIX OF 1-, 3- AND 5-YEAR-OLD SAMPLE GROUPS Parameter (year) F value (F 0.05 = 3.14) Average SD Average SD Average SD Alcoholic degree (%, v/v) Total acid (g/l) Brix (%)

6 344 H. YU ET AL. and 5-year-old sample groups are significantly different. The tendency was in accordance with that obtained by other authors (Xu et al. 2003). The decrement of alcohol was for the oxidation, esterification and volatilization reactions during aging. Part of alcohol was oxidized into acetic acid; another part was turned into acetic acid through esterification reaction; and the other volatilized through the mud seal (lid) out of the ceramic pottery. The reason for the increase in total acid was that alcohol was oxidized into aldehyde, then into acid. The possible reason for the increase of Brix was that alcohol and some higher alcohols were turned into organic acid salt through oxidation reaction during aging. Aside from changes in the three components, the concentration of amino acid nitrogen decreased gradually while that of esters, total dissolved matter and metal salts increased. Meanwhile, changes in acidity, alcohol and metal salts caused coagulation of proteide, tannin, metal icons and oligosaccharides in wine body (Xu et al. 2003). Spectral Analysis Figure 1 shows the average spectra of the 1-, 3- and 5-year-old sample groups of Tapai brand in the calibration set. The main features of the spectra are absorption bands at 1455 and 1900~1950 nm, which are related to the first Absorbance Absorbance Wavelength (nm) Wavelength (nm) FIG. 1. AVERAGE SPECTRA OF 1-, 3- AND 5-YEAR-OLD SAMPLE GROUPS OF TAPAI BRAND IN THE CALIBRATION SET

7 CLASSIFICATION OF CHINESE RICE WINE USING NIR SPECTROSCOPY Arbitrary units years old 3 years old 5 years old Arbitrary units Wavelength (nm) Wavelength (nm) FIG. 2. SECOND DERIVATIVE SPECTRA OF 1-, 3- AND 5-YEAR-OLD SAMPLE GROUPS OF TAPAI BRAND IN THE CALIBRATION SET overtone of the O H stretch of H 2 O and a combination of stretch and deformation of the O H group in H 2 O and ethanol, respectively (Maeda et al. 1999; Buning-Pfaue 2003). The strongest absorption band at 1900~1950 nm is also in the NIR spectrum of pure water. The absorption band at 2266 nm is likely related to C H combinations and O H stretch overtones, and absorption band at 2305 nm is related to C H overtones. Absorption bands at 1450, 1790 and 2266 nm were also reported to be associated with sucrose, fructose and glucose in fruit juices (Rambla et al. 1997; Roggp et al. 2004). The spectral differences between the three sample groups can be detected from 2200 to 2500 nm. Transformed by second derivative, the spectra were inverted. Then, the absorption bands, which corresponded to each major absorption band in the original spectra, became narrow valleys. The spectral differences between the 1-, 3- and 5-year-old sample groups are at 1902, 1958, 2250, 2286, 2335 and 2365 nm, which are shown in Fig. 2. From Fig. 2, it can also be seen that the difference between the 1- and 5-year-old sample groups is the greatest, followed by the difference between the 1- and 3-year-old sample groups; and that the average spectra of the 3-year-old sample group is very close to that of the 5-year-old sample group.

8 346 H. YU ET AL. Chemometric Analysis PCA. To investigate the basis for the spectral discrimination, PCA eigenvectors were analyzed (Fig. 3). The first four principal components (PCs) account for 99.58% of the variation in the spectra. PC1 explains 89.55% of the total variance, and the highest eigenvector is found at 2066 nm, which is associated with the combination of stretch and deformation of the O H group, followed by that found at 1495 nm, which is associated with NH 2 group of nitrogen compounds. PC2 explains 6.10% of the variation, and the highest eigenvector is found at 1885 nm, which associated with the combination of stretch and deformation of the O H group in H 2 O, followed by that found at 1400 nm, which is associated with first overtone of the O H stretch of H 2 O. DA. Calibration Analysis. The calibration results of the DA model developed on raw spectra are better than that on the second derivative spectra (Table 3). From Table 3, it is shown that only one sample is misclassified in the DA model developed on raw spectra. While there are 11 samples incorrectly classified in the DA model developed on second derivative spectra, the percentage of samples correctly classified is 78.4%. The reason for the better performance of raw spectra than that of second derivative spectra is that DA Absorbance PC1 PC PC4 PC Wavelength (nm) FIG. 3. EIGENVECTORS FOR PRINCIPAL COMPONENT ANALYSIS ON THE CALIBRATION SET PC, principal component.

9 CLASSIFICATION OF CHINESE RICE WINE USING NIR SPECTROSCOPY 347 TABLE 3. DISCRIMINANT ANALYSIS RESULTS FOR CALIBRATION SET BASED ON RAW AND SECOND DERIVATIVE SPECTRA OF CHINESE RICE WINE SAMPLES Data type Marked age (years) Number of samples Number of samples misclassified % of samples correctly classified Raw spectra Second derivative spectra year old 5 years old Distance to 5 years Distance to 1 year FIG. 4. PAIRWISE DISTANCE PLOT FOR 1- AND 5-YEAR-OLD SAMPLE GROUPS IN THE CALIBRATION SET with raw spectra describes 100% of the variability, while DA with second derivative describes 87.3% (the data, 100 and 87.3%, are obtained from the calibration results of the two models). Thus, raw spectra were used for calibration and validation analysis. The results were not in agreement with the literature for Australian white wine variety analysis (Cozzolino et al. 2003), in which second derivative spectra together with discriminant partial least squares developed the optimum calibration results. The pairwise distance plots for the 1- and 5-year-old sample groups as well as those for the 3- and 5-year-old ones are shown in Figs. 4 and 5, respectively. They show graphically the Mahalanobis distance between each standard and the two classes that are selected for the x- and y-axes of the plot.

10 348 H. YU ET AL years old 5 years old Distance to Distance to 3 FIG. 5. PAIRWISE DISTANCE PLOT FOR 3- AND 5-YEAR-OLD SAMPLE GROUPS IN THE CALIBRATION SET The calibration standards for the class that is selected for the x-axis of the pairwise distance plot are similar to each other; the data points for those standards will be clustered in the upper left corner of the plot. Similarly, the data points for the standards selected for the y-axis will be clustered in the lower right corner of the plot. The greater the distance between the two clusters in the plot, the greater the difference between the corresponding class and the other class. From Figs. 4 and 5, it is shown that the 1-, 3- and 5-year-old sample groups can be clearly classified. Additionally, the distance between the 1- and 5-year-old sample groups is the greatest; thus, the calibration results for the Chinese rice wine samples with marked ages of 1 and 5 years are the best. In the calibration analysis, one Kuaijishan brand sample with a marked age of 3 years is classified in the 5-year-old sample group, which is the sample in the circle on the diagonal line in Fig. 5. The Mahalanobis distance of 5-year old is , a little lower than that of 3-year old (1.3841); then, it is assigned to 5 years. Comparing the reference data of the sample with the average data listed in Table 2, it can be obtained that alcoholic degree and total acid of the sample are between those of 3 and 5 years and close to 5 years, and Brix is higher than that of 5 years. Then, the marked age of the sample is considered as a little higher than 3 years. Therefore, it is concluded that the sample is correctly classified. To sum up, there is no sample misclassified in the calibration analysis, and then the percentage of samples correctly classified is 100%. The ability of the NIR spectroscopy to discriminate or identify wine is based on the vibrational responses of chemical bonds to NIR radiation. It is

11 CLASSIFICATION OF CHINESE RICE WINE USING NIR SPECTROSCOPY 349 TABLE 4. DISCRIMINANT ANALYSIS RESULTS FOR VALIDATION SET BASED ON RAW SPECTRA OF CHINESE RICE WINE SAMPLES Commercial label Marked age (years) Number of samples Number of samples misclassified Tapai Kuaijishan TABLE 5. REFERENCE DATA OF ALCOHOLIC DEGREE, TOTAL ACID AND BRIX AND CALCULATED AGE OF SAMPLES MISCLASSIFIED IN THE VALIDATION SET Number Commercial label Marked age (years) Calculated age (years) Alcoholic degree (%, v/v) Total acid (g/l) Brix (%) 1 Tapai Kuaijishan Kuaijishan Kuaijishan probable that the higher the variability between the sample types in those chemical entities, which respond in these regions of the spectra, the better the accuracy of the model (Cozzolino et al. 2003). From Table 2, it is shown that the major components (alcoholic degree, total acid and Brix) of Chinese rice wine with different marked ages are significantly different. Most of the changes were related to the primary structural components of organic molecules (C H, N H and O H bonds), which provided the necessary information for classification of Chinese rice wine with different marked ages by NIR spectroscopy. Meanwhile, the spectral difference between the sample groups with different marked age (1, 3 and 5 years) can be detected from Figs. 1 and 2, correspondingly. Therefore, NIR spectroscopy can be used to classify Chinese rice wine samples with different marked ages. Validation Analysis. Among the total of 18 samples for validation analysis, there were four samples misclassified (Table 4). The reference data for alcoholic degree, total acid and Brix, and calculated age of the misclassified samples are listed in Table 5. For sample 1, the values of alcoholic degree and total acid are between the average values of 1 and 3 years (Table 2), respectively; and the value of Brix is lower than that of 1 year; therefore, it is concluded that the sample is incorrectly classified. For sample 2, the values of

12 350 H. YU ET AL. alcoholic degree and Brix are between the average values of 3 and 5 years, respectively; and the value of total acid is higher than that of 5 years; therefore, it is concluded that the sample is wrongly labeled and correctly classified. Following the same rule, it can be obtained that samples 3 and 4 are both wrongly labeled and correctly classified. To sum up, there is one sample misclassified in the validation analysis, and then the percentage of samples correctly classified is 94.4%. For the 1-, 3- and 5-year-old sample groups, the percentage of samples correctly classified was 88.3, 100 and 100%, respectively. The results of this study indicated that NIR spectroscopy offered the possibility to classify wines with different marked ages without the need for costly and laborious chemical and sensory analyses. However, only a limited number of samples and wine varieties were used in this study; therefore, further research is needed. CONCLUSION The applicability of NIR spectroscopy to classify Chinese rice wine with different marked age (1, 3 and 5 years) is presented in this work. The percentages of samples correctly classified in calibration and validation analysis were 100 and 94.4%, respectively. The results indicated that NIR spectroscopy together with DA is a powerful tool for discrimination between Chinese rice wine samples with different marked age. Further studies are needed to extend to other marked age and varieties of Chinese rice wine. ACKNOWLEDGMENTS The authors gratefully acknowledge the financial support provided by the Program for New Century Excellent Talents in University (No. NCET ) and the Research Fund for the Doctoral Program of Higher Education (NO ). REFERENCES BAXTER, J.M., CREWS, M.E., DENNIS, J., GOODALL, I. and ANDER- SON, D The determination of the authenticity of wine from its trace element composition. Food Chem. 60, BUNING-PFAUE, H Analysis of water in food by near infrared spectroscopy. Food Chem. 82,

13 CLASSIFICATION OF CHINESE RICE WINE USING NIR SPECTROSCOPY 351 CAVINATO, A.G., MAYES, D.M., GE, Z.H. and CALLIS, J.B Noninvasive method for monitoring ethanol in fermentation processes using fiber-optic near- infrared spectroscopy. Anal. Chem. 62, COZZOLINO, D., KWIATKNOWSKI, M.J., PARKER, M., CYNKAR, W.U., DAMBERGS, R.G., GISHEN, M. and HEDERICH, M.J Prediction of phenolic compounds in red wine fermentations by visible and near infrared spectroscopy. Anal. Chim. Acta 513, COZZOLINO, D., SMYTH, H.E. and GISHEN, M Feasibility study on the use of visible and near-infrared spectroscopy together with chemometrics to discriminate between commercial white wines of different varietal origins. J. Agric. Food Chem. 51, DAMBERGS, R.G., KAMBOURIS, A., FRANCIS, I.L. and GISHEN, M Rapid analysis of methanol in grape-derived distillation products using near-infrared transmission spectroscopy. J. Agric. Food Chem. 50, DUGO, G., PERA, L.L., PELLICANO, T.M., BELLA, G.D. and D IMPERIO, M Determination of some inorganic anions and heavy metals in D.O.C. Golden and Amber Marsala wines: statistical study of the influence of ageing period, color and sugar content. Food Chem. 91, GARCIA-JARES, C., GARCIA-MARTIN, S. and CELA-TORRIJOS, R Analysis of some high compounds of wine by means of purge and cold trapping injector capillary gas chromatography. Application to the differentiation of Rias Baixas Spanish white wines. J. Agric. Food Chem. 43, GISHEN, M. and DAMBERGS, R.G Some preliminary trials in the application of scanning near-infrared spectroscopy (NIRS) for determining the compositional quality of grape, wine and spirit. Australian Grapegrower and Winemaker 47, KALLITHRAKA, S., ARVANITOYANNIS, I.S., KEFALAS, P., EL-ZAJOULL, A., SOUFLEROS, E. and PSARRA, E Instrumental and sensory analysis of Greek wines; implementation of principal component analysis (PCA) for classification according to geographical origin. Food Chem. 73, LEORNARDO, S.M., FLAVIA, C.C.O., PAULO, A.Z.S. and JOEL, C.R Determination of ethanol in fuel ethanol and beverages by Fourier transform (FT)-near infrared and FT-Raman spectrometries. Analytica Chimica Acta 493, MAEDA, H., WANG, Y., OZAKI, Y., SUZUKI, M., CZARNECKI, M.A. and IWAHASHI, M A near-infrared study of hydrogen bonds in alcohols comparison of chemometrics and spectroscopic analysis. Chemometr Intell Lab. 45,

14 352 H. YU ET AL. MAYEN, M., BARON, R., MAID, J. and MEDIAN, M Changes in phenolic compounds during accelerated browning in white wines from cv. Pedro Ximenez and cv. Baladi grapes. Food Chem. 58, OSBORNE, B.G., FEARN, T. and HINDLE, P.H Practical NIR Spectroscopy with Applications in Food and Beverage Analysis, 2nd Ed., Longman Scientific and Technical, Harlow, Essex, UK RAMBLA, F.J., GARRIGUES, S. and GUARDIA, M PLS-NIR determination of total sugar, glucose, fructose and sucrose in aqueous solutions of fruit juices. Anal. Chim. Acta 334, REBOLO, S., PENA, R.M., LATORRE, M.J., GARCIA, S., BOTANA, A.M. and HERRERO, C Characterisation of Galician (NW Spain) Riberira Sacra wines using pattern recognition analysis. Anal. Chim. Acta 417, ROGGP, Y., DUPONCHEL, L. and HUVENNE, J.P Quality evaluation of sugar beet by near-infrared spectroscopy. J. Agric. Food Chem. 52, SAENZ-LOPEZ, R., FERNANDEZ-ZURBANO, P. and TENA, M.T Analysis of aged wine pigments by capillary zone electrophoresis. J. Chromatogr., A 1052, SAIZ-ABAJO, M.J., GONZALEZ-SAIZ, J.M. and PIZARRO, C Classification of wine and alcohol vinegar samples based on near-infrared spectroscopy. Feasibility study on the detection of adulterated vinegar samples. J. Agric. Food Chem. 52(25), SOUFLEROS, E., BARRIOS, M.L. and BERTRAND, A Correlation between the content of biogenic amines and other wine compounds. Am. J. Enol. Vitic. 49, SOUFLEROS, E.H., BOULOUMPASI, E., TSARCHOPOULOS, C. and BILIADERIS, C.G Primary amino acid profiles of Greek white wines and their use in classification according to variety, origin and vintage. Food Chem. 80, URBANO-CUADRADO, M., LUQUE DE CASTRO, M.D., PEREZ-JUAN, P.M., GARCIA-OLMO, J. and GOMEZ-NIETO, M.A Near infrared reflectance spectroscopy and multivariate analysis in enology determination or screening of fifteen parameters in different types of wines. Anal. Chim. Acta 52, XU, R.N., BAO, Z.D., PAN, X.X. and HU, P.X Maturization of Chinese rice wine. Liquor Making 30(2), YU, H.Y., YING, Y.B., FU, X.P. and LU, H.S Quality determination of Chinese rice wine based on Fourier transform near infrared spectroscopy. J. Near Infrared Spec. 14(1),