Method for tracing producing area of cyperus esculentus by adopting infrared combined chemometrics method

Abstract

According to the method for tracing the cyperus esculentus producing area by adopting an infrared combined chemometrics method, an optimal classification and identification model and optimal model parameters for tracing the cyperus esculentus producing area are obtained, the cyperus esculentus producing area can be accurately and rapidly identified, complicated sample treatment is not needed, the analysis efficiency is high, the producing area identification is accurate, and the method is suitable for large-scale popularization and application. The method is suitable for large-scale popularization and application.

Description

Technical Field

[0001] The invention relates to the technical field of crop origin tracing, and in particular to a method for tracing the origin of cyperus oleifera by using an infrared combined with a chemometric method. Background Art

[0002] Cyperus oleiferus, also known as tiger nut, iron water chestnut, underground walnut and ginseng bean, is a non-genetically modified economic crop that integrates grain, oil, animal husbandry, feed, medicine and greening ornamental. The main varieties are large round grain, small round grain, large long grain and small long grain. Its origin is in Africa and the countries or regions along the Mediterranean coast. Now it is widely planted in Xinjiang, Heilongjiang, Beijing, Hebei, Hunan, Shandong, Sichuan and other places in China. Cyperus oleiferus has multiple uses and high added value. First, it is a high-quality oil crop with an oil content of up to 20% to 36%. It is one of the varieties with the highest per-acre yield of vegetable oil among all herbaceous plants. Secondly, the underground tubers of Cyperus oleiferus are rich in nutrients such as oil, starch, sugar, protein, dietary fiber, etc., and the above-ground stems and leaves can be used as high-quality forage for livestock and poultry. Cyperus oleiferus has high application value and is currently widely used in oil pressing, wine making, sugar making, beverages and other food fields.

[0003] There are significant differences in the nutritional content and functional development of cyperus juncea from different origins. At present, the cyperus juncea circulating in the market is relatively complex, with products of different varieties, origins, and production methods mixed together, which seriously affects the quality of cyperus juncea and the prospects of its application industry.

[0004] CN114594090A discloses a method for tracing the origin of seabuckthorn fruit. The method comprises the following steps: digesting freeze-dried powder of seabuckthorn fruit from multiple different origins to prepare test samples, and determining the element contents of Ca, Mg, K, Cu, Zn, Fe, Mn, P, Se, Ni, Mo, Co, Rb, Sr, Ba, Al, and B in the samples; using IBM SPSS 25.0 software, performing one-way analysis of variance, screening out mineral elements with significant differences (P < 0.05) between different origins, and then discriminating them, screening out index elements and obtaining function coefficients of the origins corresponding to the index elements, and establishing a discriminant model for different origins; determining the element contents of the index elements of the seabuckthorn fruit to be traced, and substituting them into the discriminant model for different origins, and the origin corresponding to the maximum mineral element content value is the origin of the seabuckthorn fruit to be identified.

[0005] CN116381085A discloses a mung bean origin tracing method based on characteristic differential metabolites. The tracing method detects mung bean samples from different origins based on non-targeted metabolomics technology combined with ultra-high performance liquid chromatography-time-of-flight mass spectrometry, obtains metabolite data of mung bean samples from different origins, and uses chemometric methods to screen characteristic differential metabolites of mung bean samples from different origins; based on the screened characteristic differential metabolites, a mung bean origin tracing classification model is constructed through machine learning to achieve accurate tracing of the mung bean origin.

[0006] However, the above-mentioned crop origin traceability method is complicated to operate, requires expensive testing equipment, and is not suitable for tracing the origin of cyperus oleifera.

[0007] Therefore, it is of great significance to develop a method that can quickly and accurately identify the geographical origin of cyperus rotundus to ensure food safety. Summary of the invention

[0008] In view of the problems existing in the prior art, the present invention provides a method for tracing the origin of cyperus oleifera by using infrared combined with chemometric methods. By conducting chemometric research on infrared spectra, the best classification and recognition model for origin tracing and the optimal model parameters are obtained, thereby achieving the purpose of simply, quickly and accurately identifying the origin of cyperus oleifera.

[0009] To achieve this object, the present invention adopts the following technical solutions:

[0010] The present invention provides a method for tracing the origin of Cyperus oleifera by using an infrared combined with a chemometric method, the method comprising the following steps:

[0011] (1) collecting infrared spectrum data of cyperus oleifera from different origins and dividing the data into a training set and a prediction set; the infrared spectrum data includes near infrared spectrum data or mid infrared spectrum data;

[0012] (2) Preprocessing of all infrared spectral data after baseline correction and smoothing;

[0013] (3) Using chemometric software, the training set was operated under different classification and recognition models. The recognition rate of the training set was calculated based on the model output results and the comparison with the actual origin of cyperus oleifera. The model with the highest recognition rate was selected as the best classification and recognition model for origin traceability, and then the prediction set was used for verification.

[0014] (4) The training set is used to optimize the model parameters of the origin traceability optimal classification and identification model to obtain the optimal model parameters, and then the prediction set is used to calculate under the origin traceability optimal classification and identification model with the optimal model parameters, and the function is used to count the data with correct origin identification, thereby obtaining the origin identification rate of Cyperus oleifera in the prediction set, thereby realizing the origin traceability of Cyperus oleifera.

[0015] The method for tracing the origin of cyperus oleiferus by using infrared combined with chemometrics method described in the present invention does not need to use high-end and expensive testing instruments. The infrared spectrometer most commonly used in crop research is used to collect infrared spectrum data of cyperus oleiferus from different origins, which is divided into a training set and a prediction set, and the infrared spectrum data is simply preprocessed; then, the chemometrics software is used to perform model calculation on the training set under different classification and recognition models to obtain the best classification and recognition model for origin tracing; then, the model parameters of the best classification and recognition model for origin tracing are optimized to obtain the optimal model parameters, and finally the model is used to realize the origin tracing of cyperus oleiferus. The method described in the present invention is simple and fast to operate, and computer software is used to perform model calculation and optimization, so that the origin of cyperus oleiferus can be identified quickly, accurately and at low cost, and is suitable for large-scale promotion and application.

[0016] The present invention does not specifically limit the method for collecting infrared spectrum data of Cyperus oleifera from different origins, and conventional operation methods in the art may be used.

[0017] Preferably, the infrared spectrum data of cyperus oleifera from different origins in step (1) are randomly sampled and divided into a training set and a prediction set at a ratio of 2:1. The training set is used to test and train various classification and recognition models and optimize the parameters of the models, and the prediction set is used to verify whether the best classification and recognition model and the optimal model parameters obtained from the training set are appropriate, so as to ensure that the correct origin of cyperus oleifera is obtained by using chemometrics software analysis.

[0018] Preferably, the wavelength range of the near infrared spectroscopy data in step (1) is 4000 to 10000 cm -1 , for example, it can be 4000cm -1 、4500cm -1 、5000cm -1 、6000cm -1 、7000cm -1 , 9000cm -1 or 10000cm -1 The above range is not limited to the listed values, and other values ​​not listed within the above range are also applicable.

[0019] Preferably, the wavelength range of the mid-infrared spectral data is 400-4000 cm -1 , for example, it can be 400cm -1 , 500cm -1 , 1000cm -1 、1500cm -1 , 2000cm -1 、3000cm -1 or 4000cm -1The above range is not limited to the listed values, and other values ​​not listed within the above range are also applicable.

[0020] The method of the present invention preferentially selects infrared spectrum data of cyperus oleifera from a relatively far origin in space as a training set, which can effectively reduce the overlap of spectral information and achieve the purpose of model simplification.

[0021] Preferably, it is characterized in that the preprocessing method in step (2) includes any one of multivariate scatter correction, multi-quantity standardization or multi-quantity standardization coupled detrending algorithm.

[0022] Preferably, the chemometrics software in step (3) includes chempattern.

[0023] Preferably, the classification recognition model in step (3) includes any one of chemometric K nearest neighbor method (KNN), orthogonal partial least squares discriminant analysis (OPLS-DA), partial least squares discriminant analysis (PLS-DA), canonical correlation analysis, cluster independent soft model or support vector machine (SVM).

[0024] The present invention adopts the above-mentioned various existing classification and recognition models to perform chemometric model operations on infrared spectrum data, which is conducive to obtaining the best classification and recognition model for origin tracing, and then realizing accurate tracing of the origin of cyperus oleifera.

[0025] Preferably, the model parameters in step (4) include sample distance, data preprocessing UV scaling, Pareto scaling, automatic scaling, homogenization and centering.

[0026] Preferably, the sample distance comprises Euclidean distance, squared Euclidean distance or block distance.

[0027] The Euclidean distance measures the straight-line distance between two points in an m-dimensional space.

[0028] In n-dimensional space, two points (x1, x2, ..., x n ) and (y1,y2,...,y n ) is:

[0029]

[0030] The squared Euclidean distance is the squared form of the Euclidean distance.

[0031] In n-dimensional space, two points (x1, x2, ..., x n ) and (y1,y2,...,y n ) is:

[0032] d 2=(x2-x1) 2 +(y2-y1) 2 +…+(x n -y n ) 2

[0033] The block distance is a method of measuring the distance between two points in a multidimensional space by calculating the sum of the absolute differences between the two points in each dimension. The formula for block distance is:

[0034] d AB =|x2-x1|+|y2-y1|+…+|n2-n1|

[0035] Preferably, the optimal model parameters in step (4) are model parameters whose recognition rate of the origin traceability best classification recognition model reaches more than 95%, for example, it can be 95%, 95.5%, 96%, 97%, 98%, 98.5% or 99%, etc., but are not limited to the listed values, and other unlisted values ​​within the numerical range are also applicable.

[0036] Preferably, the statistics in step (4) use the EXACT function and the COUNTIFS function.

[0037] As a preferred technical solution of the present invention, the method comprises the following steps:

[0038] (1) collecting infrared spectrum data of cyperus oleifera from different origins, and randomly sampling and dividing the data into a training set and a prediction set at a ratio of 2:1; the infrared spectrum data includes near-infrared spectrum data or mid-infrared spectrum data;

[0039] The wavelength range of the near infrared spectral data is 4000~10000cm -1 The wavelength range of the mid-infrared spectral data is 400-4000cm -1 .

[0040] (2) Preprocessing of all infrared spectral data after baseline correction and smoothing;

[0041] The preprocessing method includes any one of multivariate scatter correction, multi-quantity standardization or multi-quantity standardization coupled detrending algorithm;

[0042] (3) The chemometrics software ChemPattern was used to perform model calculations on the training set under different classification and recognition models. The recognition rate of the training set was calculated based on the model output results and the comparison with the actual origin of Cyperus oleifera. The model with the highest recognition rate was selected as the best classification and recognition model for origin traceability, and then the prediction set was used for verification.

[0043] The classification recognition model includes any one of chemometric K nearest neighbor method (KNN), orthogonal partial least squares discriminant analysis (OPLS-DA), partial least squares discriminant analysis (PLS-DA), canonical correlation analysis, cluster independent soft model or support vector machine (SVM);

[0044] (4) The training set is used to optimize the model parameters of the origin tracing best classification and identification model to obtain the optimal model parameters, and then the prediction set is used to calculate under the origin tracing best classification and identification model with the optimal model parameters, and the EXACT function and COUNTIFS function are used to count the data with correct origin identification, thereby obtaining the origin identification rate of Cyperus oleifera in the prediction set, and realizing the origin tracing of Cyperus oleifera;

[0045] The model parameters include sample distance, data preprocessing UV scaling, Pareto scaling, automatic scaling, homogenization and centering; the sample distance includes Euclidean distance, squared Euclidean distance or block distance; the optimal model parameters are model parameters for which the recognition rate of the origin traceability best classification recognition model reaches more than 95%.

[0046] Compared with the prior art, the present invention has at least the following beneficial effects:

[0047] The method for tracing the origin of cyperus oleifera by using infrared combined with chemometric methods provided by the present invention achieves accurate and rapid identification of the origin of cyperus oleifera through data optimization of infrared spectroscopy and chemometric methods. It is a non-destructive analysis, does not require complicated sample processing, has high analysis efficiency, and can accurately identify the origin. It provides a reference method for improving the accuracy of tracing the origin of cyperus oleifera, thereby ensuring food safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 This is the original near-infrared spectrum of the cyperus oleifera sample from Hebei Province in Example 1 of the present invention.

[0049] Figure 2 This is the MSC graph of 8 oil-saving sedge seeds pretreated by near-infrared spectroscopy in Example 1 of the present invention.

[0050] Figure 3 This is the result diagram of the near infrared tracing of the origin of Cyperus oleifera in Example 1 of the present invention.

[0051] Figure 4 This is the original mid-infrared infrared spectrum of the Xinjiang cyperus erinaceus sample in Example 2 of the present invention.

[0052] Figure 5 This is the MSC graph of 8 oil-saving cyperus rotundus pretreated by mid-infrared spectroscopy in Example 2 of the present invention.

[0053] Figure 6This is the mid-infrared tracing result diagram of the origin of Cyperus oleifera in Example 2 of the present invention. DETAILED DESCRIPTION

[0054] The technical solution of the present invention is further described below with reference to the accompanying drawings and through specific implementation methods.

[0055] The present invention is further described in detail below. However, the following examples are only simplified examples of the present invention and do not represent or limit the scope of protection of the present invention. The scope of protection of the present invention shall be subject to the claims.

[0056] Example 1

[0057] This embodiment provides a method for tracing the origin of Cyperus oleifera by using infrared combined with chemometrics, the method comprising the following steps:

[0058] (1) A total of 585 samples of Cyperus oleifera were collected from eight provinces (or autonomous regions), including Shandong, Hunan, Jilin, Inner Mongolia, Ningxia, Yunnan, Xinjiang, and Hebei. All samples were numbered in chronological order of collection, washed, dried in the sun, crushed at room temperature, passed through a 50-mesh sieve, and stored at 20°C. The spectra of all samples were collected using a Fourier near-infrared spectrometer, with a spectral acquisition parameter setting of 32 cm resolution. -1 , scan times 64, wavelength range 4000~10000cm -1 After the equipment was turned on and preheated for 30 minutes, the detector (InAsTEC) gain was 80.36. The cyperus rotundus samples were evenly placed in 20 mL vials and then placed on a rotating tray. The compactness of the samples in the tray was kept basically consistent. Each sample was scanned and measured 3 times. The original near-infrared spectrum of the cyperus rotundus sample in Hebei Province is shown in the figure below. Figure 1 As shown; the data set is divided into training set and prediction set according to the random sampling ratio of 2:1;

[0059] (2) After baseline correction and smoothing, all infrared spectral data were preprocessed; the preprocessing method was multivariate scattering correction; MSC graphs of the preprocessed near infrared spectra of 8 cyperus rotundus were shown in Figure 2. Figure 2 As shown;

[0060] (3) The chemometrics software ChemPattern was used to perform model operations on the training set under different classification and recognition models. The recognition rate of the training set was calculated based on the model output results and the comparison with the actual origin of cyperus oleifera. The metrology K-nearest neighbor method (KNN) with the highest recognition rate was selected as the best classification and recognition model for origin traceability, and then the prediction set was used for verification.

[0061] Specifically, representative samples were set in the chemometrics software, and sample points that did not meet the identification requirements were eliminated. The results are shown in Table 1, with 346 samples in the training set and 172 samples in the prediction set.

[0062] Table 1

[0063]

[0064] As shown in Table 1, in the KNN mode, UV scaling, leave-one-out cross validation, and the sample distance are block distances, the recognition rate of the training set can reach 91.4%, and the model needs to be further optimized.

[0065] The 172-sample prediction set of near-infrared spectrometer was used, combined with chempattern chemometrics software, and spectral preprocessing was used, including data cleaning, normalization, and standardization. These steps can remove noise and reduce the impact of different dimensions and data distribution differences on the model. Select appropriate data processing, set sample distance and K value, start software training and prediction, and allocate them into one category according to the closest point in space. The overall prediction rate can reach up to 96.7%, and the province with the highest recognition rate has achieved a prediction accuracy of 100%, see Table 2.

[0066] Table 2

[0067]

[0068]

[0069] It can be concluded from Table 2 that the optimal model has the highest prediction recognition rate for Yunnan Province. It can be seen from the table that the representative samples of the training set are 308, 281 and 293, the predicted samples are close to the representative samples, and the recognition rate can reach 100%.

[0070] (4) The training set is used to optimize the model parameters of the KNN model to obtain the optimal model parameters. Then, the prediction set is used to calculate under the KNN model with the optimal model parameters, and the EXACT function and COUNTIFS function are used to count the data with correct origin identification, thereby obtaining the origin identification rate of the cyperus rotundus in the prediction set, and realizing the origin tracing of the cyperus rotundus. The near infrared cyperus rotundus origin tracing result in this embodiment is shown in FIG. Figure 3 As shown;

[0071] The optimal model parameters include that the sample distance is the block distance, the data preprocessing is UV scaling, the dimension reduction is 80% variance explanation, and the above optimal model parameters are model parameters for which the recognition rate of the KNN model reaches more than 95%.

[0072] In this embodiment, near-infrared spectra of cyperus oleifera samples from 8 provinces were collected, and model training was performed through chemometric methods to obtain the best classification and recognition model for origin traceability: the metrological K nearest neighbor method (KNN) and the corresponding optimal model parameters, which completed the rapid and accurate identification of the origin of each cyperus oleifera sample in the prediction set.

[0073] Example 2

[0074] This embodiment provides a method for tracing the origin of Cyperus oleifera by using infrared combined with chemometrics, the method comprising the following steps:

[0075] (1) A total of 585 samples of Cyperus oleifera were collected from eight provinces (or autonomous regions), including Shandong, Hunan, Jilin, Inner Mongolia, Ningxia, Yunnan, Xinjiang, and Hebei. All samples were numbered in chronological order of collection, washed, dried in the sun, crushed at room temperature, passed through a 50-mesh sieve, and stored at -20°C. The spectra of the Cyperus oleifera samples were collected using a Fourier mid-infrared spectrometer. The Cyperus oleifera samples and KBr were pressed into tablets at a mass ratio of 1:100. The spectrum collection parameters were set as follows: resolution 4 cm -1 , scan times 64, wavelength range 400~4000cm -1 , aperture: medium resolution, sample chamber: iZ10 / right AEM. After the equipment is turned on and preheated for 30 minutes, the detector DTGSKBr collects spectral fingerprint data. The original mid-infrared spectrum of the Xinjiang cyperus sample is shown in the figure below. Figure 4 As shown; the data set is divided into training set and prediction set according to the random sampling ratio of 2:1;

[0076] (2) After baseline correction and smoothing, all infrared spectral data were preprocessed; the preprocessing method was multivariate scattering correction; MSC graphs of the preprocessed mid-infrared spectra of 8 cyperus rotundus were shown in Figure 2. Figure 5 As shown;

[0077] (3) The chemometrics software Chempattern was used to perform model operations on the training set under different classification and recognition models. The recognition rate of the training set was calculated based on the model output results and the comparison with the actual origin of cyperus oleifera. The chemometrics K-nearest neighbor method (KNN) with the highest recognition rate was selected as the best classification and recognition model for origin traceability, and then the prediction set was used for verification.

[0078] Specifically, representative samples were set in the chemometrics software, and sample points that did not meet the identification requirements were eliminated. The results are shown in Table 3, with 347 samples in the training set and 172 samples in the prediction set.

[0079] Table 3

[0080]

[0081] As shown in Table 3, in the KNN mode, automatic scaling, leave-one-out cross validation, and Euclidean distance are used, the recognition rate of the training set can reach 96.8%, and the model is reliable.

[0082] The 172 sample prediction set of the mid-infrared spectrometer was used in combination with the chempattern chemometrics software, and spectral preprocessing was performed, including data cleaning, normalization, and standardization. These steps can remove noise and reduce the impact of different dimensions and data distribution differences on the model. Select appropriate data processing, set the sample distance and K value, start software training and prediction, and classify them into one category based on the closest point in space. The overall prediction rate can reach up to 96.8%, and the prediction accuracy of some samples in the province with the highest recognition rate can reach 100%, see Table 4.

[0083] Table 4

[0084]

[0085]

[0086] It can be concluded from Table 4 that the optimal model has the highest prediction recognition rate for Xinjiang Uygur Autonomous Region. It can be seen from the table that the representative samples of some training sets are 52, 51, 55, 28, and 42. The predicted samples are close to the representative samples, and the recognition rate can reach 100%.

[0087] (4) The training set is used to optimize the model parameters of the origin tracing optimal classification and recognition model to obtain the optimal model parameters, and then the prediction set is used to calculate under the origin tracing optimal classification and recognition model with the optimal model parameters, and the EXACT function and COUNTIFS function are used to perform statistics on the data with correct origin identification, and then the origin identification rate of the cyperus rotundus in the prediction set is obtained to achieve the origin tracing of the cyperus rotundus; the infrared cyperus rotundus origin tracing result in this embodiment is shown in FIG. Figure 6 As shown;

[0088] The optimal model parameters include that the sample distance is the squared Euclidean distance and the data preprocessing is homogenization. The optimal model parameters are model parameters that enable the recognition rate of the best classification recognition model for origin traceability to reach more than 95%.

[0089] In this example, mid-infrared spectra of cyperus oleifera samples from eight provinces were collected, and model training was performed using a chemometrics method to obtain the best classification and recognition model for origin traceability: the metrology K nearest neighbor method (KNN) and the corresponding optimal model parameters, which completed the rapid and accurate identification of the origin of each cyperus oleifera sample in the prediction set.

[0090] In summary, the method for tracing the origin of cyperus oleifera using infrared combined with chemometrics provided by the present invention can accurately and quickly identify the origin of cyperus oleifera, does not require complicated sample processing, has high analysis efficiency, accurately identifies the origin, and is suitable for large-scale promotion and application.

[0091] The applicant declares that the above is only a specific implementation mode of the present invention, but the protection scope of the present invention is not limited thereto. Those skilled in the art should understand that any changes or substitutions that can be easily thought of by those skilled in the art within the technical scope disclosed by the present invention shall fall within the protection scope and disclosure scope of the present invention.

Claims

1. A method for tracing the origin of Cyperus oleifera by using infrared combined with chemometrics, characterized in that: The method comprises the following steps: (1) collecting infrared spectrum data of cyperus oleifera from different origins and dividing the data into a training set and a prediction set; the infrared spectrum data includes near infrared spectrum data or mid infrared spectrum data; (2) Preprocessing of all infrared spectral data after baseline correction and smoothing; (3) Using chemometric software, the training set was operated under different classification and recognition models. The recognition rate of the training set was calculated based on the model output results and the comparison with the actual origin of cyperus oleifera. The model with the highest recognition rate was selected as the best classification and recognition model for origin traceability, and then the prediction set was used for verification. (4) The training set is used to optimize the model parameters of the origin traceability optimal classification and identification model to obtain the optimal model parameters, and then the prediction set is used to calculate under the origin traceability optimal classification and identification model with the optimal model parameters, and the function is used to count the data with correct origin identification, thereby obtaining the origin identification rate of Cyperus oleifera in the prediction set, thereby realizing the origin traceability of Cyperus oleifera.

2. The method according to claim 1, characterized in that In step (1), the infrared spectrum data of cyperus oleifera from different origins are randomly sampled and divided into a training set and a prediction set in a ratio of 2:

1.

3. The method according to claim 1 or 2, characterized in that: The wavelength range of the near infrared spectroscopy data in step (1) is 4000-10000cm -1 ; The wavelength range of the mid-infrared spectral data is 400-4000 cm -1 .

4. The method according to any one of claims 1 to 3, characterized in that: The preprocessing method in step (2) includes any one of multivariate scatter correction, multi-quantity standardization or multi-quantity standardization coupled detrending algorithm.

5. The method according to any one of claims 1 to 4, characterized in that: The chemometrics software in step (3) includes chempattern.

6. The method according to any one of claims 1 to 5, characterized in that: The classification and recognition model in step (3) includes any one of the chemometric K nearest neighbor method, orthogonal partial least squares discriminant analysis, partial least squares discriminant analysis, canonical correlation analysis, cluster independent soft model or support vector machine.

7. The method according to any one of claims 1 to 6, characterized in that: The model parameters in step (4) include sample distance, data preprocessing UV scaling, Pareto scaling, automatic scaling, homogenization and centering.

8. The method according to claim 7, characterized in that The sample distance includes Euclidean distance, squared Euclidean distance or block distance.

9. The method according to any one of claims 1 to 8, characterized in that: The optimal model parameters in step (4) are model parameters that enable the recognition rate of the origin tracing best classification recognition model to reach more than 95%.

10. The method according to any one of claims 1 to 9, characterized in that: The statistics in step (4) use the EXACT function and the COUNTIFS function.

Images