Quantitative analysis method for volatile oil content in bighead atractylodes rhizome

Abstract

The quantitative analysis method comprises the following steps: acquiring near infrared spectrum data and volatile oil content data of a bighead atractylodes rhizome sample, and constructing a training set and a test set; preprocessing the near infrared spectrum data of the bighead atractylodes rhizome sample; screening characteristic wave numbers from the preprocessed near infrared spectrum data, wherein the screening method is a collaborative interval partial least square method or a competitive adaptive reweighted sampling variable selection method; taking the near infrared spectrum data corresponding to the screened characteristic wave number as input, taking the corresponding volatile oil content as output, and training the PLS model by adopting the training set and the test set to obtain a volatile oil content prediction model; and acquiring near infrared spectrum data of a to-be-detected atractylodes macrocephala sample, performing pretreatment, and inputting the near infrared spectrum data corresponding to the characteristic wave number into the volatile oil content prediction model to obtain a volatile oil content prediction value of the to-be-detected atractylodes macrocephala sample. The quantitative analysis method has the advantages of simplicity, convenience, rapidness, no damage and the like.

Description

Technical Field

[0001] This invention relates to the field of quality analysis technology of Chinese medicinal materials, and in particular to a quantitative analysis method for the volatile oil content in Atractylodes macrocephala. Background Technology

[0002] Atractylodes macrocephala Koidz., the dried rhizome of the plant (Atractylodes macrocephala Koidz.), belongs to the Asteraceae family. It is bitter, sweet, and warm in nature, and enters the spleen and stomach meridians. It has the effects of strengthening the spleen and replenishing qi, drying dampness and promoting diuresis, and dispelling wind and cold. It can be used for symptoms such as spleen deficiency with poor appetite, abdominal distension and diarrhea, phlegm retention with dizziness and palpitations, edema, spontaneous sweating, and threatened abortion. One of the main active components of Atractylodes macrocephala is its volatile oil. Modern pharmacological studies have shown that the volatile oil of Atractylodes macrocephala has various biological activities, including anti-inflammatory, anti-tumor, antibacterial, antioxidant, and immunomodulatory activities. The volatile oil content is often used as a standard for evaluating the quality of Atractylodes macrocephala.

[0003] Currently, the conventional analytical methods for Atractylodes macrocephala volatile oil mainly include steam distillation, gas chromatography (GC), gas chromatography-mass spectrometry (GC-MS), and high performance liquid chromatography (HPLC). For example, Liu Shaowen (Liu Shaowen. Comparative study on the determination of volatile oil content in different Atractylodes macrocephala species [J]. China Contemporary Medicine, 2010, 17(24):2.DOI:10.3969 / j.issn.1674-4721.2010.24.027.) used GC-MS to detect the volatile oil content in Atractylodes macrocephala and explored the variation law of volatile oil content in Atractylodes macrocephala harvested at different years and periods. Wang Feng et al. (Wang Feng, Cai Guangming, Guo Huiling. Determination of atractylone content in Atractylodes macrocephala volatile oil by high performance liquid chromatography [J]. Central South Pharmacy, 6.3(2008):3.) used high performance liquid chromatography to detect the volatile oil content in Atractylodes macrocephala. While these methods are highly accurate and reproducible, they suffer from complex sample pretreatment, are time-consuming and costly, and can damage samples. They are not suitable for rapid and non-destructive testing of large batches of samples, nor can they meet the needs for rapid on-site testing in the production, processing and circulation of Chinese medicinal materials.

[0004] Near-infrared spectroscopy is a technique that measures chemical structure by detecting molecular bonds (such as CH, NH, and OH) in the near-infrared region. It quantifies or qualitatively analyzes a sample based on its absorption, reflection, or transmission of near-infrared light. It has advantages such as high universality, rapid detection, accurate results, and non-destructive nature. Therefore, near-infrared spectroscopy has great potential in predicting the content of components in traditional Chinese medicine. Summary of the Invention

[0005] This invention provides a quantitative analysis method for the volatile oil content in Atractylodes macrocephala, which has the advantages of being simple, rapid, and non-destructive.

[0006] The technical solution of the present invention is as follows: A quantitative analysis method for the volatile oil content in Atractylodes macrocephala includes the following steps: (1) Obtain near-infrared spectral data and volatile oil content data of Atractylodes macrocephala samples, and construct training and testing sets; (2) The near-infrared spectral data of Atractylodes macrocephala samples were preprocessed using at least one of the following methods: first derivative processing, multiplicative scattering correction, and standard normal variable processing. (3) Select characteristic wavenumbers from the preprocessed near-infrared spectral data. The selection method is either the cooperative interval partial least squares method or the competitive adaptive reweighted sampling variable selection method. (4) The near-infrared spectral data corresponding to the selected characteristic wavenumbers are used as input, and the corresponding volatile oil content is used as output. The PLS model is trained using the training set and the test set to obtain the volatile oil content prediction model. (5) Obtain the near-infrared spectral data of the Atractylodes macrocephala sample to be tested, preprocess it according to the method in step (2), and input the near-infrared spectral data corresponding to the characteristic wavenumber into the volatile oil content prediction model to obtain the predicted value of the volatile oil content of the Atractylodes macrocephala sample to be tested.

[0007] Preferably, the preparation method of Atractylodes macrocephala sample includes: grinding the dried Atractylodes macrocephala sample into powder and passing it through a 50-mesh sieve.

[0008] Preferably, the near-infrared spectral data of Atractylodes macrocephala samples include those in the range of 4000~10000 cm⁻¹. -1 Within the wavenumber range of 5~10 cm -1 Spectral data is acquired at high resolution and corrected using the atmospheric background spectrum as a reference.

[0009] In the range of 4000~10000 cm -1 In the wavenumber range, the near-infrared spectrum showed several strong spectral peaks corresponding to the vibrations of specific groups, including combined vibrations of CH and CC (4000 cm⁻¹). -1 The second-order harmonic vibration of the carbonyl group (5350 cm⁻¹) -1 The first harmonic vibrations of OH and NH (6900 cm⁻¹) -1 ), and CH's stretching, asymmetric stretching vibration and asymmetric deformation vibration (7200 cm). -1 This can make the prediction results more accurate.

[0010] Further preferred, the near-infrared spectrum of each sample was measured three times, and the average value was used for variable analysis.

[0011] Preferably, the volatile oil content of Atractylodes macrocephala was determined using the volatile oil determination method in the Pharmacopoeia of the People's Republic of China (2025 edition, Part IV).

[0012] Preferably, in step (2), the preprocessing is multiplicative scattering correction processing.

[0013] Preferably, in step (3), the selection of characteristic wavenumbers using the cooperative interval partial least squares method includes: dividing the full spectrum into several intervals, then selecting a combination of multiple intervals to establish a PLS model, using root mean square error as the evaluation index, and selecting the optimal characteristic wavenumbers.

[0014] Further optimization involves dividing the full spectrum into 30 intervals, and then selecting the combination of the 11th, 22nd, and 24th intervals to establish a PLS model.

[0015] Preferably, in step (3), the selection of feature wavenumbers using the competitive adaptive reweighted sampling variable selection method includes: (i) Use 80% of the random samples as the calibration set for the PLS model; (ii) Use the exponentially decreasing function to remove wavelengths with smaller regression coefficients; (iii) Use the adaptive reweighted sampling method to screen out wavelengths with larger regression coefficients; (iv) Select the subset of wavenumbers with the smallest root mean square error of cross-validation values ​​in the PLS model to obtain the characteristic wavenumbers.

[0016] Further preferably, in step (3), the selected characteristic wavenumbers include: 4265.768 cm⁻¹ -1 4288.91 cm -1 4292.767 cm -1 4296.624 cm -1 4396.904 cm -1 4400.761 cm -1 4404.618 cm -1 5218.431cm -1 5295.569 cm -1 7559.589 cm -1 7563.446 cm -1 8315.548 cm -1 8330.976 cm -1 8334.833 cm -1 8338.689 cm -1 9761.898 cm -1 .

[0017] Compared with the prior art, the beneficial effects of the present invention are as follows: This invention provides a method for rapid analysis of volatile oil content in Atractylodes macrocephala based on near-infrared spectroscopy. The sample preparation process is convenient and non-destructive, with no risk of chemical reagent contamination, and the detection is rapid and the results are accurate. Attached Figure Description

[0018] Figure 1 This is the near-infrared spectrum of a sample of Atractylodes macrocephala.

[0019] Figure 2 This is a scatter plot of the training and test sets of Atractylodes macrocephala volatile oil under the Si-PLS model, where A is the training set and B is the test set. Detailed Implementation

[0020] The present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be noted that the embodiments described below are intended to facilitate the understanding of the present invention and do not limit it in any way.

[0021] Unless otherwise specified, the methods used in the following embodiments are all existing techniques in the art. Unless otherwise specified, the experimental materials used in the following embodiments were all purchased from conventional biochemical reagent companies.

[0022] Example 1: Near-infrared quantitative model of volatile oil in Atractylodes macrocephala based on different pretreatment methods I. Collection of Near-Infrared Spectra and Volatile Oil Content of Atractylodes macrocephala Samples (1) Prepare quantitative samples of Atractylodes macrocephala powder, dry the sample and grind it through a 50-mesh sieve.

[0023] (2) Near-infrared spectra were acquired using a Fourier transform near-infrared spectrometer in diffuse reflectance mode. Approximately 2 g of Atractylodes macrocephala powder sample was evenly packed into a round sample cup and collected at 4000–10000 cm⁻¹. -1 Within the wavenumber range at 8 cm -1 Spectral acquisition was performed at high resolution. Thirty-two scans were performed, with calibration using the atmospheric background spectrum as a reference. The spectrum of each sample was measured three times, and the average value was used for variable analysis.

[0024] Figure 1 The near-infrared spectrum of the Atractylodes macrocephala sample is shown. Within this region, the near-infrared spectrum exhibits several strong spectral peaks corresponding to vibrations of specific functional groups, including combined vibrations of CH and CC (4000 cm⁻¹). -1 The second-order harmonic vibration of the carbonyl group (5350 cm⁻¹) -1 The first harmonic vibrations of OH and NH (6900 cm⁻¹) -1 ), and CH's stretching, asymmetric stretching vibration and asymmetric deformation vibration (7200 cm). -1 ).

[0025] (3) The content of volatile oil in Atractylodes macrocephala powder in step (1) was determined by referring to the volatile oil determination method (General Rule 2204) of the Pharmacopoeia of the People's Republic of China (2025 edition, Part IV).

[0026] II. Spectral Data Preprocessing Six signal preprocessing techniques were employed, including first derivative, multiplicative scattering correction, standard normal variables, and combinations of two different methods, to mitigate baseline drift and amplify subtle spectral differences between samples.

[0027] III. Establishment of PLS ​​Quantitative Analysis Model Using spectral data with different preprocessing steps as input and Atractylodes macrocephala volatile oil content data as output, a PLS quantitative analysis model was established, and the performance of each model was evaluated.

[0028] The performance of the model is evaluated based on six metrics: (i) the correlation coefficient (RC) in the training set. 2 (ii) Correlation coefficients (RP) in the test set 2 (iii) Root mean square error (RMSEC) of the training set; (iv) Root mean square error (RMSEP) of the test set; (v) Relative analytical error (RPDC) of the training set; (vi) Relative analytical error (RPDP) of the test set. 2 A value close to 1 and an RMSE close to 0 indicate that the constructed model has good accuracy and robustness. RPD is used to evaluate predictive ability; RPD > 2 indicates good prediction.

[0029] The relevant evaluation calculation formulas for the model are as follows: (1) (2) (3) in: , These represent the standard value and the predicted value of the sample, respectively. denoted by , where is the mean of the components of the sample set; N is the number of observations; and SD is the standard deviation. When the sample set Y is the training set, the result is the parameter of the training set; when the sample set Y is the test set, the result is the parameter of the training set.

[0030] Table 1 shows the performance parameters of each PLS quantitative analysis model for volatile oils established using spectral data with different preprocessing methods. Among all methods, the optimal preprocessing method is multiplicative scattering correction (RP). 2 =0.8547, RMSEV=0.1040, PRD=2.6774).

[0031] Table 1 Comparison of the performance of different spectral preprocessing methods on the PLS quantitative model

[0032] Example 2: Constructing a near-infrared quantitative model for volatile oils in Atractylodes macrocephala based on different spectral characteristic wavenumber methods I. Screening of characteristic wavenumbers in Atractylodes macrocephala spectra using the co-located partial least squares variable selection method The preprocessed full spectrum of Atractylodes macrocephala was divided into 10, 20, or 30 intervals. Then, combinations of 2, 3, or 4 intervals within each interval were selected to establish a PLS model (Si-PLS). Table 2 shows the corresponding characteristic wavenumber selection and its root mean square error. The optimal characteristic wavenumbers obtained by applying the synergistic interval partial least squares variable selection method were set to 30 intervals, with intervals 11, 22, and 24 selected.

[0033] Table 2. Comparison of model performance for screening characteristic wavenumbers of Atractylodes macrocephala spectra using the synergistic interval partial least squares variable selection method.

[0034] II. Using the competitive adaptive reweighted sampling variable selection method to screen the characteristic wavenumbers of Atractylodes macrocephala spectra After optimal preprocessing, 80% of the random samples were used as the calibration set for the PLS regression model. Wavelengths with smaller regression coefficients were removed using an exponentially decreasing function. An adaptive reweighted sampling method was used to filter out wavelengths with larger regression coefficients. Finally, the subset of wavelengths with the smallest root mean square error of cross-validation values ​​in the PLS model (CRAS-PLS) was selected. A competitive adaptive reweighted sampling variable selection method was applied to the PLS model construction, resulting in 16 characteristic wavenumbers, 4265.768 cm⁻¹. -1 4288.91 cm -1 4292.767 cm -1 4296.624 cm -1 4396.904 cm -1 4400.761 cm -1 4404.618 cm -1 5218.431 cm -1 5295.569 cm -1 7559.589 cm -1 7563.446 cm -1 8315.548 cm -1 8330.976cm -1 8334.833 cm -1 8338.689 cm -1 9761.898 cm -1 .

[0035] III. Construction of Quantitative Models Similar to Example 1, the performance of PLS ​​quantitative analysis models established by different characteristic wavenumber screening methods is compared.

[0036] Table 3 shows the performance parameters of each PLS quantitative analysis model for Atractylodes macrocephala volatile oil established using spectral data selected by the characteristic wavenumber screening method. The optimal characteristic wavenumber screening method is the synergistic interval partial least squares variable selection method (RP). 2 =0.8894, RMSEV=0.0875, PRD=3.0695).

[0037] Table 3 Comparison of the performance of different spectral characteristic wavenumber screening methods on PLS quantitative models

[0038] Figure 2 The scatter plots of the training and test sets of Atractylodes macrocephala volatile oil under the Si-PLS model are shown: A is the training set and B is the test set.

[0039] The embodiments described above provide a detailed explanation of the technical solutions and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the present invention. Any modifications, additions, and equivalent substitutions made within the scope of the principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A quantitative analysis method for the volatile oil content in Atractylodes macrocephala, characterized in that, Includes the following steps: (1) Obtain near-infrared spectral data and volatile oil content data of Atractylodes macrocephala samples, and construct training and testing sets; (2) The near-infrared spectral data of Atractylodes macrocephala samples were preprocessed using at least one of the following methods: first derivative processing, multiplicative scattering correction, and standard normal variable processing. (3) Select characteristic wavenumbers from the preprocessed near-infrared spectral data. The selection method is either the cooperative interval partial least squares method or the competitive adaptive reweighted sampling variable selection method. (4) The near-infrared spectral data corresponding to the selected characteristic wavenumbers are used as input, and the corresponding volatile oil content is used as output. The PLS model is trained using the training set and the test set to obtain the volatile oil content prediction model. (5) Obtain the near-infrared spectral data of the Atractylodes macrocephala sample to be tested, preprocess it according to the method in step (2), and input the near-infrared spectral data corresponding to the characteristic wavenumber into the volatile oil content prediction model to obtain the predicted value of the volatile oil content of the Atractylodes macrocephala sample to be tested.

2. The method for quantitative analysis of volatile oil content in Atractylodes macrocephala according to claim 1, characterized in that, The preparation method of Atractylodes macrocephala sample includes: grinding the dried Atractylodes macrocephala sample into powder and passing it through a 50-mesh sieve.

3. The method for quantitative analysis of volatile oil content in Atractylodes macrocephala according to claim 1, characterized in that, Near-infrared spectral data of Atractylodes macrocephala samples were obtained in the range of 4000–10000 cm⁻¹. -1 Within the wavenumber range of 5~10 cm -1 Spectral data is acquired at high resolution and corrected using the atmospheric background spectrum as a reference.

4. The method for quantitative analysis of volatile oil content in Atractylodes macrocephala according to claim 3, characterized in that, Near-infrared spectra of each sample were measured three times, and the average values ​​were used for variable analysis.

5. The method for quantitative analysis of volatile oil content in Atractylodes macrocephala according to claim 1, characterized in that, The volatile oil content of Atractylodes macrocephala was determined using the volatile oil determination method in the Pharmacopoeia of the People's Republic of China (2025 edition, Part IV).

6. The method for quantitative analysis of volatile oil content in Atractylodes macrocephala according to claim 1, characterized in that, In step (2), the preprocessing is multiplicative scattering correction processing.

7. The method for quantitative analysis of volatile oil content in Atractylodes macrocephala according to claim 1, characterized in that, In step (3), the selection of characteristic wavenumbers using the cooperative interval partial least squares method includes: dividing the full spectrum into several intervals, then selecting a combination of multiple intervals to establish a PLS model, using root mean square error as the evaluation index, and selecting the optimal characteristic wavenumbers.

8. The method for quantitative analysis of volatile oil content in Atractylodes macrocephala according to claim 7, characterized in that, The full spectrum is divided into 30 intervals, and then the combination of the 11th, 22nd and 24th intervals is selected to build the PLS model.

9. The method for quantitative analysis of volatile oil content in Atractylodes macrocephala according to claim 1, characterized in that, In step (3), the selection of feature wavenumbers using the competitive adaptive reweighted sampling variable selection method includes: (i) Use 80% of the random samples as the calibration set for the PLS model; (ii) Use the exponentially decreasing function to remove wavelengths with smaller regression coefficients; (iii) Use the adaptive reweighted sampling method to screen out wavelengths with larger regression coefficients; (iv) Select the subset of wavenumbers with the smallest root mean square error of cross-validation values ​​in the PLS model to obtain the characteristic wavenumbers.

10. The method for quantitative analysis of volatile oil content in Atractylodes macrocephala according to claim 1 or 9, characterized in that, In step (3), the selected characteristic wavenumbers include: 4265.768 cm⁻¹ -1 4288.91 cm -1 4292.767 cm -1 4296.624 cm -1 4396.904 cm -1 4400.761 cm -1 4404.618 cm -1 5218.431 cm -1 5295.569 cm -1 7559.589cm -1 7563.446 cm -1 8315.548 cm -1 8330.976 cm -1 8334.833 cm -1 8338.689 cm -1 9761.898 cm -1 .

Images