Construction and combined evaluation analysis method of thin-layer chromatography fingerprint of compound donkey-hide gelatin pulp

Abstract

The invention discloses a compound donkey-hide gelatin pulp thin-layer chromatography fingerprint construction and combination evaluation analysis method, and belongs to the technical field of traditional Chinese medicine quality control. The compound colla corii asini pulp thin-layer chromatography is a thin-layer chromatography fingerprint spectrum of the compound colla corii asini pulp or an intermediate in a process preparation process; the construction of the thin-layer chromatography fingerprint spectrum is based on the color of a thin-layer chromatography sample strip image and the gray scale of spots, and a pixel-intensity matrix and a ratio shift value-gray scale matrix are formed to comprehensively reflect the thin-layer chromatography characteristics of the sample; according to the combined evaluation method of the thin-layer chromatography fingerprint spectrum, a method of combining machine learning with information entropy and entropy is adopted, ginsenoside Rg1, codonopsis pilosula alkynol and other components are screened and identified as key components influencing the compound donkey-hide gelatin pulp and transmission, and the method has the characteristics of accuracy and high efficiency, is wide in application prospect and can provide a thought for establishment of quality standards of traditional Chinese medicine compounds.

Description

Technical Field

[0001] The present invention relates to the technical field of quality control of traditional Chinese medicines, and in particular to the construction of a thin-layer chromatographic fingerprint of a compound donkey-hide gelatin slurry and an analysis method thereof. Background Art

[0002] Thin layer chromatography (TLC) is a commonly used method for the identification, quality evaluation, and control of traditional Chinese medicines (TCMs). It can obtain chemical profiles from complex samples and is characterized by simplicity, speed, economy, and sensitivity. Fingerprinting, on the other hand, allows for a more intuitive assessment of the overall chemical profile of a sample and the extraction of macroscopic regularities within the sample, making it a common method for the study of TCM quality standards. Data sources are diverse, including time-series spectra or images from high-performance liquid chromatography, mass spectrometry, and thin-layer chromatography. Traditional TLC analysis relies heavily on empirical experience and reference materials to analyze image results, lacking objective and precise analytical evaluation methods. TLC images feature complex colored bands. By analyzing the band color and grayscale of the spots, a characteristic chemical profile information set can be generated for each band. This allows for the construction of a TLC fingerprint, enabling the analysis and evaluation of similarities and differences between samples from a holistic perspective.

[0003] Traditional methods of analyzing samples by identifying the primary spots of reference substances or reference medicinal materials lack an overall understanding of the bands, potentially overlooking spots with lower grayscale values, i.e., components with relatively low content. Furthermore, efforts to extract effective key information from thin-layer chromatography fingerprints for sample analysis and evaluation have limited to separate evaluations of TLC band grayscale values ​​or color (RGB) using principal component analysis (PCA) and hierarchical cluster analysis (HCA) to classify samples, lacking comprehensive analysis and evaluation of band grayscale values ​​and color. Currently, no research has used machine learning combined with information entropy and entropy methods to comprehensively evaluate the color or spots of TLC images.

[0004] Compound donkey-hide gelatin paste (FEJ) is a traditional Chinese medicine made from donkey-hide gelatin, red ginseng, Rehmannia glutinosa, Codonopsis pilosula and hawthorn through modern processing and refinement. It is widely used in clinical practice and can be used for symptoms such as qi and blood deficiency, dizziness, palpitations, insomnia, and anemia. The "Chinese Pharmacopoeia" (2020 edition) requires the identification items of red ginseng and Codonopsis pilosula in FEJ to use ginsenotriol and Codonopsis pilosula as control materials, respectively. This invention is based on the team's early ideas for the overall quality control of compound donkey-hide gelatin paste (application number: CN202111041118.5) to further study the thin-layer chromatography of red ginseng and Codonopsis pilosula in compound donkey-hide gelatin paste and process preparation intermediates, and identify the components suitable for quality control, which provides a reference for the improvement of the quality standards of compound donkey-hide gelatin paste, and at the same time provides ideas for the establishment of quality control and quality standards of traditional Chinese medicine compounds. Summary of the Invention

[0005] In view of this, the object of the present invention is to provide a method for constructing a thin-layer chromatographic fingerprint of a compound donkey-hide gelatin slurry and conducting a combined evaluation and analysis.

[0006] In order to achieve the above-mentioned object of the invention, the present invention provides the following technical solutions:

[0007] The present invention provides a method for constructing a thin-layer chromatographic fingerprint of a compound donkey-hide gelatin slurry, comprising the following steps:

[0008] (1) Obtaining thin layer chromatographic images of different batches of compound donkey-hide gelatin syrup or intermediates during the process of preparation;

[0009] (2) Generation of “pixel-intensity” matrix and “ratio-shift value-grayscale” matrix;

[0010] Furthermore, the image acquisition in step (1) is a thin layer chromatography image acquired under natural light and subjected to desaturation and denoising processing by median filtering the image;

[0011] Furthermore, the "pixel-intensity" matrix in step (2) is obtained by longitudinally cropping the thin layer chromatography image, and the RGB channels in the longitudinally cropped strip are extracted using ImageJ software, and the "Plot Profile" function is used to generate the corresponding intensity of the pixels, and the RGB intensity value is inverted by subtracting the intensity value of each color channel from the maximum intensity value, thereby forming a matrix;

[0012] Furthermore, the “ratio shift value-grayscale” matrix described in step (2) is obtained by horizontally cropping the thin layer chromatography image, and the “Measure” function of the ImageJ software is used to measure the grayscale of the spots of the horizontally cropped strips, and the shift distance is measured to calculate the ratio shift value (Rf), thereby forming.

[0013] The present invention also provides a method for combined evaluation and analysis of thin layer chromatography fingerprints, comprising the following steps:

[0014] (1) Using machine learning to filter “pixel-intensity” matrix features;

[0015] (2) Comprehensive evaluation of information entropy and entropy value into “ratio shift value-grayscale” matrix;

[0016] Furthermore, the method of using a machine learning method to screen features in step (1) is to use a machine learning method to screen features on a "pixel-intensity" matrix, and select pixels with weights and VIP values ​​greater than 0 and their corresponding intensities as features;

[0017] Furthermore, the machine learning method described in step (1) is random forest;

[0018] Furthermore, the information entropy calculation formula in step (2) is:

[0019]

[0020] Among them, X i Indicates the percentage of the grayscale corresponding to each feature in the grayscale corresponding to all features;

[0021] Furthermore, the entropy calculation formula in step (2) is:

[0022]

[0023] Among them, X i Indicates the percentage of the grayscale corresponding to each feature in the grayscale corresponding to all features.

[0024] The present invention has the following beneficial effects:

[0025] The construction and combined evaluation and analysis method of the thin-layer chromatography fingerprint of the compound donkey-hide gelatin slurry provided by the present invention can be applied to the thin-layer chromatography image processing process of the finished compound donkey-hide gelatin slurry and the intermediates in the process of process preparation and transfer, accurately identify and screen the characteristic spots in the compound ingredient system, and evaluate the importance of the characteristic spots to the sample, providing new ideas and methods for the research of quality markers of traditional Chinese medicine compounds and the establishment of a quality evaluation system. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following is a brief introduction to the drawings required for use in the embodiments. It should be understood that the following drawings only illustrate certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other relevant drawings can be obtained based on these drawings without paying any creative work.

[0027] Figure 1 The thin layer chromatograms for identifying red ginseng and codonopsis pilosula in different batches of compound donkey-hide gelatin solution;

[0028] Among them, 1 to 15 are different batches of compound donkey-hide gelatin slurry, with batch numbers: 2112031, 2201007, 2112029, 2205047, 2205044, 2208033, 2204014, 2206011, 2204017, 2203018, 2204016, 2207024, 2207008, 2203010, and 2207004.

[0029] Figure 2 To transmit the thin layer chromatograms for identifying red ginseng and codonopsis pilosula for the process of “medicinal materials-decoction pieces-intermediates-prescription preparations” for compound donkey-hide gelatin paste;

[0030] Among them, 1: Ginseng; 2: Red Ginseng; 3: Codonopsis pilosula medicinal material; 4: Codonopsis pilosula decoction pieces; 5 to 11 are intermediates of compound donkey-hide gelatin slurry, namely: the first extract, the second extract, the third extract, the liquid before adding donkey-hide gelatin, the liquid after adding donkey-hide gelatin, the liquid before filling and the liquid after sterilization; 12 to 14 are finished preparations of compound donkey-hide gelatin slurry, with batch numbers: 2112031, 2201007, 2112029 respectively.

[0031] Figure 3 To study the chemometric analysis results of thin layer chromatography fingerprint for the quality standard of compound donkey-hide gelatin solution;

[0032] A: PCA analysis of R channel; B: HCA analysis of R channel; C: PCA analysis of G channel; D: HCA analysis of G channel; E: PCA analysis of B channel; F: HCA analysis of B channel.

[0033] Figure 4 The chemometric analysis results of thin layer chromatography fingerprint for the delivery study of compound donkey-hide gelatin syrup;

[0034] A: PCA analysis of R channel; B: HCA analysis of R channel; C: PCA analysis of G channel; D: HCA analysis of G channel; E: PCA analysis of B channel; F: HCA analysis of B channel.

[0035] Figure 5 Screening the RGB channel feature pixels of TLC fingerprint for random forest;

[0036] A: Quality standard research; B: Delivery research.

[0037] Figure 6 is the entropy value result of characteristic component. DETAILED DESCRIPTION

[0038] The technical solution of the present invention is further described in detail below in conjunction with the embodiments, but the protection scope of the present invention is not limited thereto.

[0039] Example 1: Thin-layer chromatography identification of red ginseng and codonopsis pilosula in compound donkey-hide gelatin syrup and "medicinal materials-decoction pieces-intermediates-prescription preparations"

[0040] 1.1 Instrument

[0041] Ultrasonic cleaner (model KQ-250DE, Kunshan Ultrasonic Instrument Co., Ltd.);

[0042] 1.2 Materials

[0043] Compound donkey-hide gelatin paste, ginseng, red ginseng slices, codonopsis and its slices, and compound donkey-hide gelatin paste intermediates were all provided by Dong-E-E-Jiao Co., Ltd.; silica gel G thin layer plate (Qingdao Ocean Chemical Co., Ltd., batch number 20220402); red ginseng control medicinal material (China Food and Drug Inspection Institute, batch number: 121045-200604); ginsenoside Re (China Food and Drug Inspection Institute, batch number: LENA-CAAA); ginsenoside Rf (Chengdu Purifa Co., Ltd., batch number: 121045-200604); Co., Ltd., batch number: PRF9061942); ginsenoside Rh1 (Chengdu Aifa Biotechnology Co., Ltd., batch number: AF20050408); ginsenoside Rh2 (China Food and Drug Inspection Institute, batch number: 111748-202102); ginsenoside Ro (China Food and Drug Inspection Institute, batch number: 111903-202106); codonopsis pilosula alcohol (batch number 136171-87-4); other reagents not mentioned are commercially available.

[0044] 1.3 Preparation of test solution

[0045] Take 20 mL of compound donkey-hide gelatin slurry, shake and extract once with 20 mL of water-saturated n-butanol, collect the n-butanol solution, shake and extract once with 20 mL of ammonia test solution, collect the n-butanol solution, and then shake and extract 3 times with n-butanol-saturated water, 20 mL each time, combine the n-butanol solution, recover the solvent to dryness, add 1 mL of methanol to dissolve the residue, and use it as the test solution.

[0046] The preparation of the test sample solutions of the compound donkey-hide gelatin slurry intermediate, including the first extract, the second extract, the third extract, the solution before adding donkey-hide gelatin, the solution after adding donkey-hide gelatin, the solution before filling and the solution after sterilization, are all consistent with the preparation method of the compound donkey-hide gelatin slurry test sample.

[0047] The preparation method of ginseng and red ginseng test solution is consistent with the preparation method of red ginseng control medicinal material under "1.2 Selection of Control Materials" in Example 1. The preparation method of Codonopsis pilosula medicinal material and decoction piece test solution is consistent with the preparation method of Codonopsis pilosula control medicinal material under "1.4 Selection of Control Materials" in Example 1.

[0048] 1.4 Preparation of control solution

[0049] Take ginsenoside Ro, ginsenoside Re, ginsenoside Rf, ginsenoside Rh1, ginsenoside Rh2, and codonopsis quinol reference substances and add methanol to prepare solutions containing 1 mg per 1 mL. This serves as the reference solution. Separately, take 1 g of each of red ginseng and codonopsis reference medicinal materials, add 50 mL of water, heat under reflux for 45 minutes, filter, evaporate the filtrate to dryness, add 20 mL of water to the residue, and prepare the reference medicinal material solution by the same method starting from "extract once with 20 mL of water-saturated n-butanol."

[0050] 1.5 Chromatographic conditions

[0051] According to the thin layer chromatography method (General Rule 0502), 8 μL of the test solution, 1 μL of the reference solution, 2 μL of the red ginseng reference medicinal material, and 8 μL of the codonopsis reference medicinal material were respectively spotted on the same silica gel G thin layer plate, and chloroform-methanol-1% formic acid aqueous solution (6.5:2.5:0.5) was used as the developing agent. The plate was developed, removed, dried, sprayed with 10% sulfuric acid ethanol solution, heated at 105°C until the spots were clearly colored, and examined under sunlight.

[0052] Example 2: Construction of thin layer chromatography fingerprint of compound donkey-hide gelatin slurry

[0053] 3.1 Image acquisition and preprocessing

[0054] Thin-layer chromatography images were collected under natural light. The images were imported into Adobe Photoshop software, desaturated and denoised using a median filter, and then cropped and rotated horizontally and vertically to extract the corresponding images.

[0055] 3.2 Establishment of the “Pixel-Intensity” Matrix

[0056] ImageJ software was used to extract the RGB channels of the preprocessed longitudinal images, and the “Plot Profile” function was used to generate a pixel intensity matrix corresponding to each pixel in the R, G, and B channels. The RGB intensity values ​​were inverted by subtracting the intensity value of each color channel from the maximum intensity value of 255.

[0057] 3.3 Establishment of “ratio shift value-grayscale” matrix

[0058] The “Measure” function of ImageJ software was used to measure the grayscale of the spots of the transversely cropped strips, and the displacement was measured to calculate the relative displacement value (Rf), thereby forming the image.

[0059] 3.4 Establishment of thin layer chromatography fingerprints and similarity evaluation

[0060] In the thin-layer chromatographic fingerprint of compound donkey-hide glue syrup and the thin-layer chromatographic fingerprint of "medicinal materials-decoction pieces-intermediates-prescription preparations", the pixel intensity map of the R, B, and G channels of compound donkey-hide glue syrup sample S1 was used as a reference, with a data width of 1 pixel. The average method was used to generate the reference map, and peak correction and matching were performed to generate the thin-layer chromatographic fingerprint of compound donkey-hide glue syrup. Using the reference map as a reference, the cosine similarity method was used to calculate the similarity. The calculation formula (1) is as follows.

[0061]

[0062] Among them, xi represents the pixel value, and yi represents the intensity under the corresponding pixel value.

[0063] The results showed that the similarity between the R, G, and B channels of the thin-layer chromatographic fingerprint of each sample and the reference fingerprint of the compound donkey-hide gelatin syrup was in the range of 0.9551 to 0.9981, indicating that the compound donkey-hide gelatin syrup had high similarity between different batches and a stable process. In the thin-layer chromatographic fingerprint of "herbal medicine-decoction piece-intermediate-prescription preparation", the similarity between the R, G, and B channels and the reference fingerprint was in the range of 0.8383 to 0.9962, indicating that all samples had a certain degree of similarity, but it also suggested that the composition of the samples had changed to a certain extent during the transfer process, resulting in differences in the thin-layer chromatographic results.

[0064] Example 3: Combined evaluation and analysis method of thin layer chromatography fingerprint

[0065] 4.1 Chemometric analysis

[0066] To further elucidate the similarities and differences between different batches of compound donkey-hide gelatin syrup and the differences in the "herbal medicine-cooked slices-intermediate-finished product" transfer process, PCA and HCA in chemometrics were used to analyze the samples. The results of PCA and HCA showed that the samples in both the quality standard and transfer studies could be divided into three categories. There were certain differences between different batches of compound donkey-hide gelatin syrup, which may be caused by factors such as different batches of raw medicinal materials and decoction slices or process. In the "herbal medicine-cooked slices-intermediate-finished product" transfer study, it was found that the medicinal materials, decoction slices, and intermediates had certain differences with compound donkey-hide gelatin syrup, suggesting that attention should be paid to the preparation process of compound donkey-hide gelatin syrup to ensure stable and consistent quality.

[0067] 4.2 Random Forest Feature Selection

[0068] The random forest method is used to screen the features of the pixel intensity matrix of the RGB channels of the sample. The results are shown in the attached figure. Figure 5 As shown. Using reference substances, spots 1 to 5 in sample A were identified as ginsenoside Re, ginsenoside Rf, a mixture of ginsenoside Rg1, ginsenoside Rg2, and ginsenoside Rg3, ginsenoside Rh1, and codonopsis quinol. Meanwhile, spots 1 to 7 in sample B were identified as ginsenoside Re, ginsenoside Rf, a mixture of ginsenoside Rg1, ginsenoside Rg2, and ginsenoside Rg3, codonopsis quinol, ginsenoside Rh1, ginsenoside Rh2, and codonopsis quinol. Random forest feature screening revealed these components as key ingredients affecting their quality.

[0069] 4.3 Comprehensive evaluation analysis based on information entropy and entropy value

[0070] Furthermore, in order to describe the degree to which the screening characteristic components affect the quality, the grayscale values ​​of the above components are analyzed using the information entropy method. Based on the features obtained by random forest screening and the "ratio shift value-grayscale" matrix, the corresponding information entropy model is constructed. The information entropy calculation formula is as follows:

[0071]

[0072] Among them, X i Indicates the percentage of the grayscale corresponding to each feature in the grayscale corresponding to all features.

[0073] The results showed that the information entropy of the TLC fingerprints of 15 batches of compound donkey-hide gelatin syrup was 2.3219, and the information entropy of the transfer TLC fingerprint of the "herbal medicine-cooked slices-intermediates-prescription preparation" was 2.5850. Information entropy is a measure of uncertainty; greater uncertainty indicates greater disorder, meaning that the transfer TLC fingerprint contains more information about the differences between samples than the compound donkey-hide gelatin syrup TLC fingerprint, and the sample differences are greater.

[0074] In order to explore the extent to which characteristic components affect the quality and transfer process of compound donkey-hide gelatin paste, the entropy value of each component was calculated and evaluated using the entropy method. The entropy value calculation formula is as follows:

[0075]

[0076] Among them, X i Indicates the percentage of the grayscale corresponding to each feature in the grayscale corresponding to all features.

[0077] The results are as attached Figure 6 As shown in the figure, it can be seen that ginsenoside Rg1, ginsenoside Rg2, ginsenoside Rg3 and codonopsis pilosula alcohol play an important role in the quality and delivery of compound donkey-hide gelatin.

[0078] 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 principles of the present invention. These improvements and modifications should also be regarded as within the scope of protection of the present invention.

Claims

1. A method for constructing a thin-layer chromatographic fingerprint of a compound donkey-hide gelatin slurry and conducting a combined evaluation and analysis, characterized in that: The following steps are involved: (1) Obtaining thin layer chromatographic images of different batches of compound donkey-hide gelatin syrup or intermediates during the process of preparation; (2) Generate a "pixel-intensity" matrix and a "shift value-grayscale" matrix based on the stripes in the image; (3) Based on the matrix, a machine learning method is used to screen features; (4) Based on the above screening characteristics, the information entropy and entropy value are used to analyze and evaluate the thin layer chromatography strip image to obtain the evaluation results.

2. The construction and combined evaluation and analysis method of the thin layer chromatography fingerprint of the compound donkey-hide gelatin slurry according to claim 1 is characterized in that: The steps for obtaining thin-layer chromatography band images of different batches of compound donkey-hide gelatin slurry include: (1) The thin layer chromatography image of the compound donkey-hide gelatin solution was cropped horizontally and vertically respectively; (2) longitudinally cropping strips to form a "pixel-intensity" matrix; (3) The horizontal cropping strips are used to form a "ratio shift value-grayscale" matrix.

3. The construction and combined evaluation and analysis method of the thin layer chromatography fingerprint of the compound donkey-hide gelatin slurry according to claim 1 is characterized in that: The steps for feature screening using machine learning methods include: (1) Using the random forest algorithm to filter features of the "pixel-intensity" matrix of the RGB channels; (2) Pixels with weights and VIP values ​​greater than 0 and their corresponding intensities are selected as features.

4. The construction and combined evaluation and analysis method of the thin layer chromatography fingerprint of the compound donkey-hide gelatin slurry according to claim 1 is characterized in that: The evaluation steps using information entropy and entropy value analysis include: (1) Use information entropy to analyze the "ratio shift value-grayscale" matrix; (2) The entropy value was used to quantitatively evaluate the extent to which the characteristic components affected the quality of the compound donkey-hide gelatin paste and the degree to which it affected the transfer process.

5. The method of obtaining thin layer chromatography band images of different batches of compound donkey-hide gelatin slurry according to claim 2, characterized in that: (1) ImageJ software was used to extract the RGB channels in the longitudinally cropped strips. The "Plot Profile" function was used to generate the corresponding intensity of the pixels. The RGB intensity values ​​were inverted by subtracting the intensity value of each color channel from the maximum intensity value to form a "pixel-intensity" matrix. (2) The "Measure" function of ImageJ software was used to measure the grayscale of the spots in the horizontally cropped strips, and the displacement was measured to calculate the relative displacement value (Rf) and form a "ratio displacement value-grayscale" matrix.

6. The step of analyzing and evaluating using information entropy and entropy value according to claim 4, characterized in that: The calculation formulas for information entropy and entropy value are: (1) The calculation formula of information entropy is: Among them, X i Indicates the percentage of the grayscale corresponding to each feature in the grayscale corresponding to all features; (2) The formula for calculating entropy is: Among them, X i Indicates the percentage of the grayscale corresponding to each feature in the grayscale corresponding to all features.

Images