A method for identifying wine from small producing areas

Abstract

The present invention belongs to the technical field of wine detection, and discloses a method for identifying wine from small production areas, comprising the following steps: collecting grape samples, crushing and clarifying to extract juice, and making wine from the remaining grapes; determining the contents of 45 mineral elements in the obtained grape juice and wine samples; combining the determination data of grape juice and wine, applying linear discriminant analysis and multi-layer perception-artificial neural network to establish a data model, and identifying small wine production areas. The present invention clarifies the characteristics of small production areas, supplements the grape and wine data in the Diqing Plateau production area, and improves the accuracy of identifying wine from small production areas.

Description

Technical Field

[0001] The invention relates to the technical field of wine detection, in particular to a method for identifying wine from a small producing area. Background Art

[0002] In recent years, wine origin identification technology can be mainly divided into mass spectrometry, chromatography, spectroscopy and wave spectroscopy. Mass spectrometry and chromatography are usually used in combination, such as liquid chromatography-isotope ratio mass spectrometry (LC-IRMS), liquid chromatography mass spectrometry (LC-MS), gas chromatography mass spectrometry (LC-MS), etc., which can determine isotopes, non-volatile and volatile metabolites. The application of spectroscopy in wine origin identification is also increasing, including infrared spectroscopy and ultraviolet spectroscopy, which can detect metabolites in samples non-destructively. Spectroscopy is mainly based on the application of nuclear magnetic resonance hydrogen spectrum, which can determine amino acids, organic acids, alcohols, sugars, phenols, esters and organic bases in wine. A variety of wine origin tracing technologies have been established, but small production areas are usually affected by the same environment and human factors, and present similar data fingerprints in the wine, which makes it easier for samples from small production areas to be confused and their origin characteristics are ignored.

[0003] The Diqing Plateau production area in Yunnan is located on the southeastern edge of the Qinghai-Tibet Plateau and in the hinterland of the Hengduan Mountains. It is the highest altitude wine production area in my country, with an altitude range of 1700m to 3000m, and has a variety of environmental climate and geographical conditions. The production area mainly includes six small production areas, namely Benzilan Town, Yanmen Township, Yunling Township, Foshan Township, Shengping Town and Yangla Township, and has produced many high-end wines recognized by the world. In 2018, my country's first single-vineyard cultivation management standard for wine grapes was implemented in the production area, which stipulated the cultivation area, environmental conditions, harvest period and yield, etc., to form characteristic wines of small production areas, but did not provide technical guarantees for the identification of small production areas. In addition, the production area has not yet obtained my country's geographical indication product certification and small production area certification. Therefore, collecting characteristic data of the production area plays a vital role in protecting the characteristic wines of my country's high-quality production areas. Summary of the invention

[0004] The purpose of the present invention is to provide a method for identifying wine from a small producing area, so as to provide a basis for identifying wine from a small producing area.

[0005] The technical solution of the present invention is as follows:

[0006] The present invention provides a method for identifying wine from a small producing area, comprising the following steps:

[0007] Collect grape samples, crush and clarify to extract juice, and make wine with the remaining grapes;

[0008] The contents of 45 mineral elements in the obtained grape juice and wine samples were determined: 4 major elements K, Ca, Mg, Na, and 41 trace elements Li, Be, B, Ti, V, Cr, Mn, Co, Ni, Cu, Zn, As, Rb, Sr, Mo, Ag, Cd, Sb, Ba, La, Tl, Pb, Bi, U, Al, Ce, Se, Fe, Y, Pr, Nd, Sm, Eu, Gd, Tb, Dy, Ho, Er, Tm, Yb and Lu;

[0009] The measured data of grape juice and wine were combined, and linear discriminant analysis and multi-layer perception-artificial neural network were used to build a data model to identify small wine producing areas.

[0010] As an implementation mode, the grape juice and wine are digested by wet method to obtain clear liquid, and 45 mineral elements are determined by ICP-MS.

[0011] Preferably, the grape juice is added with 65 wt % nitric acid, mixed and allowed to stand, and then digested according to the following procedure: heating to 80° C., maintaining for 5 min, then heating to 120° C., maintaining for 20 min, and finally heating to 150° C. until the required liquid volume is reached.

[0012] Preferably, the wine digestion method comprises taking 5 volumes of wine and heating it at 90° C. until 2 volumes remain, adding 5 volumes of 65% HNO 3 thereto, and heating it to 120° C. until 2 volumes of liquid remain.

[0013] As an implementation method, Wilks' lambda is used as a statistic to analyze origin differences, and the leave-one-out cross-validation method is used to evaluate model performance.

[0014] As an implementation method, the multi-layer perception-artificial neural network model consists of an input layer, a hidden layer and an output layer. All data are standardized by subtracting the mean and dividing by the standard deviation. The input layer of the model is the output layer, and the number of hidden layers depends on the classification results and the amount of training data.

[0015] Preferably, the method of the present invention further comprises determining the carbon stable isotopes and oxygen stable isotopes in the grape juice and wine samples.

[0016] As an embodiment, the carbon stable isotope determination method includes placing grape juice and wine samples and an adsorbent in a tin capsule, and strictly wrapping them for standby use, and applying an element analyzer-isotope ratio mass spectrometer for determination.

[0017] As an embodiment, the oxygen stable isotope determination includes: taking grape juice and wine samples into a sampling bottle, injecting standard carbon dioxide gas to perform H2O-CO2 equilibrium in the sample, then aspirating the gas after the exchange is completed with an injection needle, and injecting it into a water balance instrument-isotope ratio mass spectrometer for determination.

[0018] Preferably, the small production areas described in the present invention are three small production areas in Benzilan Town, Deqin County, Diqing Plateau, Yunnan, Foshan Township and Yunling Township.

[0019] Compared with the prior art, the present invention has the following technical effects:

[0020] The present invention aims at the problem of identifying the origin of wine in small producing areas in my country, combines the mineral elements and isotope data of grape juice and wine, and applies multivariate analysis methods to improve the accurate identification of wine in small producing areas. The present invention clarifies the characteristics of small producing areas, supplements the grape and wine data in the Diqing Plateau producing area, and improves the accuracy of identifying wine in small producing areas. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 Box plots of stable isotopes of carbon and oxygen in grape juice and wine from three small producing areas;

[0022] Figure 2 Principal component analysis loading diagram (left) and score diagram (right) for grape juice data from three small producing areas;

[0023] Figure 3 Principal component analysis loading diagram (left) and score diagram (right) for wine data from three small producing areas;

[0024] Figure 4 Linear discriminant analysis classification results for wine data from three small producing areas;

[0025] Figure 5 Linear discriminant analysis classification results for grape juice data from three small producing areas;

[0026] Figure 6 Comprehensive discriminant analysis classification results for grape juice and wine data from three small producing areas;

[0027] Figure 7 Classification results of multi-layer perceptron artificial neural network for wine data from three small producing areas. DETAILED DESCRIPTION

[0028] The present invention provides a method for identifying wine from a small producing area, comprising the following steps:

[0029] 1) Collect grape samples, crush and clarify to extract juice, and use the remaining grapes to make wine;

[0030] The present invention preferably collects grape samples from three small grape producing areas, namely Benzilan Town, Foshan Township and Yunling Township, Deqin County, Diqing Plateau, Yunnan Province, and removes stems and crushes them before centrifugation, and takes clarified grape juice and stores it at -20°C; the remaining grapes are fermented in a small container to brew dry red wine, which is bottled in brown bottles, stored at 4°C for use, and opened after returning to room temperature before analysis.

[0031] 2) Determine the contents of 45 mineral elements in the obtained grape juice and wine samples.

[0032] As an embodiment, the grape juice and wine of the present invention are pre-treated by wet digestion to obtain a clear liquid, and 45 mineral elements are determined by ICP-MS.

[0033] As an embodiment, the grape juice is added and mixed with nitric acid and allowed to stand for heating digestion. Preferably, the concentration of nitric acid is 60-70wt%, more preferably 65%. Preferably, the volume ratio of the grape juice to the nitric acid is 1-1.5, more preferably 1. Preferably, the heating digestion procedure is to heat to 80°C-90°C, keep for 5min-7min, then heat to 100°C-120°C, keep for 15min-25min, and finally heat to 130°C-150°C to the required liquid volume. More preferably, heat to 80°C, keep for 5min, then heat to 120°C, keep for 20min, and finally heat to 150°C, until the remaining liquid is 1mL, then take it out and cool to room temperature. Transfer the digested grape juice liquid in the digestion tube to a volumetric flask, rinse the digestion tube three times with ultrapure water, and set the volume to 25mL.

[0034] As an embodiment, the wine digestion method comprises taking 5 volumes of wine and heating it at 80-90°C until 2 volumes remain, then adding 5 volumes of 60%-70% HNO3 thereto, and heating it to 100°C-120°C until the remaining liquid is 2 volumes. Specifically, 5 mL of wine is placed in a 50 mL digestion tube, heated at 90°C until 2 mL remain, then adding 5 mL of 65% HNO3 thereto, and heating it to 120°C until the remaining liquid is 2 mL.

[0035] The present invention prepares ultrapure water as a blank sample and a 500 μg / L mixed standard sample as a quality control sample according to the digestion procedures of grape juice and wine samples, wherein the mixed standard sample is a constant element standard solution with a configuration concentration of 500 μg / L.

[0036] The present invention determines the mineral elements of the grape juice and wine after the above digestion. ICP-MS is used to determine 45 kinds of mineral elements, including 4 major elements K, Ca, Mg, and Na, and 41 trace elements Li, Be, B, Ti, V, Cr, Mn, Co, Ni, Cu, Zn, As, Rb, Sr, Mo, Ag, Cd, Sb, Ba, La, Tl, Pb, Bi, U, Al, Ce, Se, Fe, Y, Pr, Nd, Sm, Eu, Gd, Tb, Dy, Ho, Er, Tm, Yb, and Lu. The range of the standard curve of major elements is 2 to 5000 μg / L, and the range of the standard curve of trace elements is 0.01 to 25 μg / L. The correlation coefficient (R 2 )≥0.999, recovery rate between 80% and 120%, relative standard deviation (RSD≤9%), if the data obtained does not meet the standard, re-measurement is required. The instrument selects helium (He) collision gas mode to eliminate Ar + , Cl - and Na + Interference, daily determination of standard solution calibration instrument drift, and rhodium (Rh) as the internal standard added to all samples to be tested, including quality control and blank. As an embodiment, the specific instrument conditions of the ICP-MS are as follows: RF power of 1100W, plasma gas flow of 15L / min, carrier gas flow of 0.94L / min, auxiliary gas flow of 1.2L / min, lens voltage of 5.5V, atomization chamber temperature of 28°C, peristaltic pump speed of 0.1r / s, and sampling depth of 8mm.

[0037] 2') Determine the carbon stable isotopes and oxygen stable isotopes in the obtained grape juice and wine samples.

[0038] As a preferred embodiment, the present invention further comprises determining the carbon stable isotopes and oxygen stable isotopes in the grape juice and wine samples. However, there is no significant difference in the determination results in the three small production areas of Benzilan Town, Foshan Township and Yunling Township. When detecting other small production areas, the determination of carbon stable isotopes and oxygen stable isotopes in grape juice and wine can be added to improve the recognition accuracy.

[0039] Carbon stable isotopes (δ 13 C) determination. Put the grape juice and wine samples and the adsorbent in a tin capsule and wrap them tightly for use. Use elemental analyzer-isotope ratio mass spectrometry (EA-IRMS) for determination. As an implementation method, high-purity helium (99.999%) is purged at 50 mL / min for 20 seconds before injection. After injection, the sample is burned in a 1000°C oxidation furnace to form a gas mixture, which is passed through a 600°C reduction furnace and then dehydrated with anhydrous magnesium perchlorate. Subsequently, the mixed gas enters a chromatographic column for separation. Finally, δ 13C. Sucrose standard substance (δ 13 C PDB =-26.03‰) more accurate laboratory standard (starch, δ 13 C PDB =-24.96‰), each sample is measured repeatedly, preferably 10 times, and the relative standard deviation does not exceed 0.1‰. As an embodiment, the IRMS instrument conditions are set to high voltage 1986V, half plate voltage -10V, focus voltage 1807V, electron voltage -80V and reverse voltage -3V. The measurement results are expressed as δ = [(R sample -R standard ) / R standard ]×1000 calculation.

[0040] Oxygen stable isotopes (δ 18 O) determination. Take grape juice and wine samples in a sampling bottle, inject standard carbon dioxide gas to balance H2O-CO2 in the sample, then draw the gas after the exchange with the injection needle and inject it into the water balance instrument-isotope ratio mass spectrometer (EQ-IRMS) for determination. The average standard ocean water (δ 18 O smow =0.00‰) calibration laboratory third-grade water, its δ 18 O smow =-7.52‰, keep the room temperature constant, preferably 23.5°C, repeat the measurement for each sample, preferably 10 times, and the relative standard deviation does not exceed 0.2‰. Specific instrument conditions and results Calculate the carbon stable isotope.

[0041] 3) Combine grape juice and wine data, and apply linear discriminant analysis (LDA) and multi-layer perception-artificial neural network (ANN-MLP) to establish a data model. The present invention uses Wilks' lambda as a statistic to analyze the differences in origin, and evaluates the model performance by leave-one-out cross validation (LOO-CV). The ANN-MLP model consists of an input layer, a hidden layer, and an output layer. All data are standardized by subtracting the mean and dividing by the standard deviation as the input layer of the ANN-MLP model. The output layer is three small production areas, and the number of hidden layers depends on the classification results and the amount of training data.

[0042] The results show that the present invention combines the data of grape juice and wine, and the accuracy rate is improved by 5.6% compared with using the data of wine alone.

[0043] The technical solution of the present invention is further described in detail below in conjunction with specific embodiments. The technical solution of the present invention includes but is not limited to the following embodiments.

[0044] Example 1

[0045] Grape samples were collected from Benzilan Town (BZL), Foshan Township (FS) and Yunling Township (YL) of Deqin County. The grapes were destemmed and crushed and then centrifuged to obtain clarified grape juice. The remaining grapes with corresponding serial numbers were destemmed and crushed, and dry red wine was brewed using the small container winemaking method.

[0046] Pretreatment for determination of mineral elements: 5 mL of grape juice sample and 5 mL of 65% nitric acid (HNO3) were added to the digestion tube and allowed to stand at room temperature for 3 hours. Then, the following procedures were followed on the hot plate: heating to 80°C, maintained for 5 minutes, then heating to 120°C, maintained for 20 minutes, and finally heating to 150°C until the remaining liquid was 1 mL, then taken out, cooled to room temperature, the liquid in the digestion tube was transferred to a volumetric flask, and the digestion tube was rinsed with ultrapure water three times and fixed to 25 mL. 5 mL of wine was placed in a 50 mL digestion tube, heated at 90°C until 2 mL remained, and then 5 mL of 65% HNO3 was added thereto, and the temperature was raised to 120°C until the remaining liquid was 2 mL. Blank (ultrapure water) and quality control samples (500 μg / L mixed standard) were prepared according to the digestion procedures for grape and wine samples.

[0047] Determination of mineral elements: ICP-MS was used to determine 45 mineral elements, including 4 major elements potassium, calcium, magnesium and sodium, and 41 trace elements (Li, Be, B, Ti, V, Cr, Mn, Co, Ni, Cu, Zn, As, Rb, Sr, Mo, Ag, Cd, Sb, Ba, La, Tl, Pb, Bi, U, Al, Ce, Se, Fe, Y, Pr, Nd, Sm, Eu, Gd, Tb, Dy, Ho, Er, Tm, Yb and Lu). The standard curve range of major elements was 2-5000 μg / L, and the standard curve range of trace elements was 0.01-25 μg / L. The correlation coefficient of the standard curve of this method (R 2 )≥0.999, recovery rate between 80% and 120%, relative standard deviation (RSD≤9%), if the data obtained does not meet the standard, re-measurement is required. The instrument selects helium (He) collision gas mode to eliminate Ar + , Cl - and Na + Interference, daily determination of standard solution to calibrate instrument drift, and rhodium (Rh) as the internal standard added to all samples to be tested, including quality control and blank. The specific instrument conditions are RF power of 1100W, plasma gas flow of 15L / min, carrier gas flow of 0.94L / min, auxiliary gas flow of 1.2L / min, lens voltage of 5.5V, spray chamber temperature of 28℃, peristaltic pump speed of 0.1r / s, and sampling depth of 8mm.

[0048] Stable isotope carbon (δ 13C) Determination: Place 2 μL of sample and a small amount of adsorbent in a tin capsule and wrap it tightly for later use. Elemental analyzer-isotope ratio mass spectrometry (EA-IRMS) is used for determination. Before injection, high-purity helium (99.999%) is purged at 50 mL / min for 20 seconds. After injection, the sample is burned in a 1000°C oxidation furnace to form a gas mixture, which is then passed through a 600°C reduction furnace and then dehydrated with anhydrous magnesium perchlorate. Subsequently, the mixed gas enters the chromatographic column for separation. Finally, δ 13 C. Sucrose standard substance (δ 13 C PDB =-26.03‰) more accurate laboratory standard (starch, δ 13 C PDB =-24.96‰), each sample was measured 10 times, and the relative standard deviation did not exceed 0.1‰. The IRMS instrument conditions were set as high voltage 1986V, half plate voltage -10V, focusing voltage 1807V, electron voltage -80V and reverse voltage -3V. The measurement results are expressed as δ = [(R sample -R standard ) / R standard ]×1000 calculation.

[0049] Stable isotope oxygen (δ 18 O) determination: Take 200 μL of sample into a sampling bottle, inject standard carbon dioxide gas to balance H2O-CO2 in the sample for 12 hours, then draw the gas after the exchange by the injection needle and inject it into the water balance instrument-isotope ratio mass spectrometer (EQ-IRMS) for determination. The average standard ocean water (δ 18 O smow =0.00‰) calibration laboratory third-grade water, its δ 18 O smow =-7.52‰, keep the room temperature at 23.5℃, repeat the measurement 10 times for each sample, and the relative standard deviation does not exceed 0.2‰. Specific instrument conditions and results calculation of stable isotope carbon.

[0050] Delta in grape juice and wine from three origins 13 C and δ 18 There was no significant difference in O Figure 1 ), compared with grape juice, the δ 13 C and δ 18 O changes relatively little and is more negative. 18 O and δ 13 C did not show good discrimination performance in identifying the origin of grape juice and wine, which may be related to the close distance and similar climate influence.

[0051] The results of variance analysis of mineral elements in the three origins (see Table 1 and Table 2) show that the elemental composition of FS and YL counties with similar soil types is closer than that of BZL. Except for rare earth elements, the contents of the remaining 16 elements in FS are the highest. As for rare earth elements, there is no significant difference in grape juices from the three origins except Er, while there are significant differences in Y, Pr, Nd, Sm and Ho in wine (p<0.05).

[0052] Table 1 Results of variance analysis of mineral elements in grape juice from three origins

[0053]

[0054]

[0055] Table 2 Results of variance analysis of mineral elements in wines from three origins

[0056]

[0057]

[0058] The principal component analysis (PCA) showed the characteristics of grape juice data in this region. The loading graph and score graph are shown in Figure 2 . PCA obtained 11 principal components with eigenvalues ​​> 1, and the variance explanation rate was 82.3%. The first principal component was positively correlated with Cd and Mg, and negatively correlated with rare earth elements Dy, Tb, Y, Nd, Ce, La, and Gd. The second principal component was positively correlated with Se, and negatively correlated with Rb, B, and Na.

[0059] The PCA results show the characteristics of the wine data in this region. The loading and score plots are shown in Figure 3 . PCA obtained 12 principal components with eigenvalues ​​> 1, and the variance explanation rate was 84.7% of the variance. The first principal component was positively correlated with K, Mg, B, Rb, Ti, Fe, Be, Cr, and Se, and negatively correlated with rare earths (Tb, Sm, Pr, Ce, Y, and La). The second principal component was positively correlated with Cd, Na, Ni, Pb, Zn, Sb, and Mn, and negatively correlated with Dy, Eu, and Mn.

[0060] Stepwise linear discriminant analysis with Wilks' lambda as the statistic was used to establish a discriminant model for the origin of grape juice and wine, and the data were visualized.

[0061] The wine origin discrimination results are shown in two discriminant functions ( Figure 4 ), the discrimination accuracy reached 75%, and the LOO-CV accuracy was 69.4%. Four samples in BZL and FS were misclassified as YL, and one sample in YL was misclassified as FS. In this model, Mn, Bi, and Ho contributed the most to origin identification.

[0062] Compared with wine, the results of grape juice origin identification ( Figure 5 ) had a high accuracy of 88.9% and LOO-CV had an accuracy of 91.7%. One sample from BZL was misclassified as FS. Two samples from YL were misclassified as BZL or FS. The grape samples from the three origins can be divided using Be, B, Cu, Zn, Sb, Pb, Bi, Pr, Dy and Er elements.

[0063] Combine grape juice with wine data ( Figure 6 ), applying the above stepwise discriminant analysis, the discrimination accuracy reached 80.6%, an increase of 5.6% compared to using wine data alone, and LOO-CV was 72.2%, an increase of 2.8%.

[0064] In order to improve the recognition accuracy of wine from small production areas, the nonlinear method ANN-MLP was further used for data analysis. 75% of the data was randomly selected as the training set and 25% as the test set. ANN-MLP was run 10 times and the average recognition accuracy of the training set was 87.7%, and the average recognition accuracy of the test set was 76.7% (Table 3). The best running results showed that the accuracy of the training set was 95.2% and the test set was 85.0% (Table 3). Figure 7 ).

[0065] Table 3 ANN-MLP data analysis results

[0066]

[0067]

[0068] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for identifying wine from a small producing area, characterized in that: The following steps are involved: Collect grape samples, crush and clarify to extract juice, and make wine with the remaining grapes; The contents of 45 mineral elements in the obtained grape juice and wine samples were determined: 4 major elements K, Ca, Mg, Na, and 41 trace elements Li, Be, B, Ti, V, Cr, Mn, Co, Ni, Cu, Zn, As, Rb, Sr, Mo, Ag, Cd, Sb, Ba, La, Tl, Pb, Bi, U, Al, Ce, Se, Fe, Y, Pr, Nd, Sm, Eu, Gd, Tb, Dy, Ho, Er, Tm, Yb and Lu; Combine the measured data of grape juice and wine, apply linear discriminant analysis and multi-layer perception-artificial neural network to build a data model, and identify small wine producing areas; After wet digestion of the grape juice and wine, 45 mineral elements were determined by ICP-MS; The grape juice is added with 60-70wt% nitric acid, mixed and allowed to stand, and then digested according to the following procedure: heating to 80-90°C, maintaining for 5-7 minutes, then heating to 100-120°C, maintaining for 15-25 minutes, and finally heating to 130-150°C until the required liquid volume is reached; The wine digestion method comprises taking 5 volumes of wine and heating it at 80-90° C. until 2 volumes remain, then adding 5 volumes of 60-70% HNO3 thereto, and heating it to 100-120° C. until 2 volumes of liquid remain.

2. The method according to claim 1, characterized in that Wilks' lambda was used as a statistic to analyze origin differences, and leave-one-out cross-validation was used to evaluate model performance.

3. The method according to claim 1, characterized in that The multi-layer perception-artificial neural network model consists of an input layer, a hidden layer and an output layer. All data are standardized by subtracting the mean and dividing by the standard deviation as the input layer of the model. The output layer is a small production area, and the number of hidden layers depends on the classification results and the amount of training data.

4. The method according to claim 1, characterized in that: The method also includes determining the carbon stable isotopes and oxygen stable isotopes in the grape juice and wine samples.

5. The method according to claim 4, characterized in that The carbon stable isotope determination method comprises placing grape juice and wine samples and an adsorbent in a tin capsule, and strictly wrapping them for standby use, and applying an element analyzer-isotope ratio mass spectrometer for determination.

6. The method according to claim 4, characterized in that The oxygen stable isotope determination comprises: taking grape juice and wine samples into a sampling bottle, injecting standard carbon dioxide gas to perform H2O-CO2 balance in the sample, then the injection needle draws the gas after the exchange is completed, and injects it into a water balance instrument-isotope ratio mass spectrometer for determination.

7. The method according to claim 1, characterized in that The small production areas mentioned are Benzilan Town, Foshan Township and Yunling Township in Deqin County, Diqing Plateau, Yunnan.

Images