Traditional Chinese medicine production place tracing method fusing element fingerprints and secondary metabolite outlines

Abstract

The invention provides a construction method, system, equipment, medium and program product of a traditional Chinese medicine origin traceability model fusing element fingerprints and secondary metabolite outlines, and correspondingly provides a traditional Chinese medicine origin traceability method, equipment, medium and program product based on the model. Relates to the field of medicinal material identification. According to the method, element fingerprints and chemical component profiles are fused according to metallomics and metabonomics, and an element fingerprint substance and secondary metabolite profile substance combined recognition mode is innovatively put forward and is used for geographical source recognition of traditional Chinese medicine with radix peucedani as a mode. And an aggregation model is innovatively adopted, and respective optimal models of elements and chemical components are further integrated, so that the classification precision is improved, and the scientificity and accuracy of traditional Chinese medicine quality evaluation are remarkably improved.

Description

Technical Field

[0001] The present invention relates to the field of medicinal material identification, and more specifically, to a method, device, medium and program product for tracing the origin of traditional Chinese medicine integrating element fingerprints and secondary metabolite profiles. Background Art

[0002] The quality and efficacy of traditional Chinese medicine are closely related to the origin of production. Accurate tracing of the characteristics of the production area can help improve quality control and effectiveness. At present, due to the depletion of wild resources, cultivated products of traditional Chinese medicine have become the main products on the market. However, the blind expansion of planting areas has led to uneven quality of traditional Chinese medicine, especially the key quality control indicators are difficult to meet the standards after cultivation in some production areas, which naturally makes it difficult to ensure the effectiveness of traditional Chinese medicine. Take Peucedanum praeruptorum as an example. With the growth of market demand, Peucedanum praeruptorum has gradually developed from wild to simulated wild, woodland and field planting. The main planting areas have expanded to Ningguo, Anhui, Hangzhou, Zhejiang, Chongqing, Sichuan, Guizhou and other places. Based on previous studies, it was found that the content of white flower peucedanum acetylcholine is closely related to the production area. Based on this, tracing the different production areas of traditional Chinese medicine is crucial to improving the scientific nature of quality evaluation and strengthening supervision.

[0003] The existing technology is relatively single-dimensional when tracing the origin of different Chinese medicines, and only constructs the element fingerprint spectrum model from the perspective of element fingerprint, such as CN201711035588.4 A method for constructing a Chinese medicinal material element fingerprint spectrum and its application in identifying the origin of Chinese medicinal materials. Or only the secondary metabolites of medicinal plants are used to reveal the origin and batch differences, but the above single-dimensional detection has been difficult to meet the current detection needs. Based on this, there is an urgent need for a more comprehensive and in-depth evaluation and traceability system for the quality of medicinal plants. Summary of the invention

[0004] The present invention aims to solve at least one of the technical problems existing in the prior art. To this end, the present invention provides a method for constructing a Chinese medicine origin tracing model integrating element fingerprints and secondary metabolite profiles, and correspondingly proposes a Chinese medicine origin tracing method, equipment, medium and program product based on the model; the present invention significantly improves the scientificity and accuracy of Chinese medicine quality evaluation by integrating element fingerprint elements and secondary metabolite profile secondary metabolite analysis multidimensional quality control strategy.

[0005] The first aspect of the present application discloses a method for constructing a Chinese medicine origin traceability model integrating element fingerprints and secondary metabolite profiles, and a method for identifying the authenticity of Chinese medicine based on multi-omics technology, the method comprising:

[0006] 101, obtaining the CPS response value of each element in the training set samples from different production areas and the first production area label corresponding to the sample;

[0007] 102, inputting the CPS response value into a first machine learning model to obtain a prediction classification result, comparing it with the first production area label, and optimizing the model according to the comparison result to obtain a metallomics model;

[0008] 103, obtaining mass spectrometry data of secondary metabolites of training set samples from different production areas and second production area labels corresponding to the samples; the second production area label is consistent with the first production area label;

[0009] 104, inputting the mass spectrometry data into a second machine learning model to obtain a prediction classification result, comparing it with the second production area label, optimizing the model according to the comparison result, and obtaining a metabolomics model;

[0010] 105. The optimal model of the metallomics model and metabolomics model in the same production area was selected, and the two optimal models were aggregated to obtain the aggregated model of the same production area as the traceability model of the production area.

[0011] In some embodiments, the first production area labels include at least two, namely production area one and production area two; the step 105 includes: selecting the best model in the metallomics model of production area one, selecting the best model in the metabolomics model of production area one, and aggregating the two best models to obtain the traceability model of production area one; selecting the best model in the metallomics model of production area two, selecting the best model in the metabolomics model of production area two, and aggregating the two best models to obtain the traceability model of production area two;

[0012] Optionally, the aggregation method of the aggregation model includes: bootstrap aggregation, boosting, stacking, mixing, etc.

[0013] In some embodiments, when the different production areas include 3, the production area label also includes production area 3; the optimal model in the metallomics model of production area 3 is selected, the optimal model in the metabolomics model of production area 3 is selected, and the two optimal models are aggregated to obtain the traceability model of production area 3; optionally, when the different production areas include 4, the production area label also includes production area 4; the optimal model in the metallomics model of production area 4 is selected, the optimal model in the metabolomics model of production area 4 is selected, and the two optimal models are aggregated to obtain the traceability model of production area 4;

[0014] Optionally, the training set samples include: Peucedanum chinense, Peucedanum purpurogenum, Saposhnikovia divaricata, etc.; when the training set samples are Peucedanum chinense, the first production area label or the second production area label includes any two or more of the following: Anhui, Zhejiang, Chongqing, Guizhou;

[0015] Optionally, the CPS response value is the response value after internal standard correction and blank background deduction;

[0016] Optionally, the elements include major elements and trace elements; trace elements include metal elements and non-metal elements;

[0017] Optionally, between 101 and 102, the CPS response value is further preprocessed, and the preprocessing includes: 1011, preliminary standardization processing of dividing the CPS response value by the sampling amount;

[0018] Optionally, after 1011, the preprocessing further includes: 1012, normalizing the CPS response values ​​of the major elements and trace elements in the training set samples from different production areas; performing dimensionality reduction processing on the normalized CPS response values; the dimensionality reduction processing method includes PCA, OPLS-DA and cluster heat map analysis;

[0019] Optionally, between 103 and 104, the step further includes performing a cleaning process on the mass spectrum data to reduce redundancy and highly correlated features, to obtain cleaned mass spectrum data;

[0020] Optionally, the cleaning process includes: 1031, taking the compound information with a response value greater than a first threshold as a basic data set, removing the molecular ion peaks close to the mass spectrum baseline and with accuracy; 1032, performing correlation removal processing on the molecular ion peaks close to the mass spectrum baseline and with accuracy that have been removed to obtain secondary metabolite mass spectrum data after removal;

[0021] In some embodiments, between 101 and 102, an oversampling process of the training set samples is also included, and the oversampling process includes: 100, copying or amplifying minority class data points to balance the number of samples of each category in the training data; the minority class data points include any one or more of the following: specific origin, special ingredient content;

[0022] Optionally, the methods used by the first machine learning model and / or the second machine learning model include: RF, KNN, Rpa, SVM, and XGB.

[0023] The second aspect of the present application discloses a method for tracing the origin of traditional Chinese medicine by integrating element fingerprints and secondary metabolite profiles, the method comprising: 201, obtaining metallomics data and plant metabolomics data of a sample to be tested;

[0024] 202, inputting the metallomics data and / or plant metabolomics data into the traceability model of different production areas for processing, and obtaining a classification result of whether the data belongs to the production area;

[0025] Optionally, the metallomics data includes CPS response values ​​of any one or more of the following characteristic elements: Mn element, S element, Mg element, B element;

[0026] Optionally, the plant metabolomics data includes mass spectrometry data of any one or more of the following chemical component peak information: peak 16696, peak 13779, peak 19774, peak 17630. Chemical components are secondary metabolite profiles (secondary metabolites), and isotopes are element fingerprints.

[0027] The third aspect of the present application discloses a method for tracing the origin of traditional Chinese medicine by integrating element fingerprints and secondary metabolite profiles, the method comprising: 301, obtaining CPS response values ​​of target characteristic elements of the sample to be tested and mass spectrum data of target peak information, the target characteristic elements comprising any one or more of the following: Mn element, S element, Mg element, B element; the target peak information comprising any one or more of the following: peak 16696, peak 13779, peak 19774, peak 17630;

[0028] 302, inputting the CPS response value of the target characteristic element and the mass spectrum data of the target peak information into the traceability model of different production areas for processing, and obtaining a classification result of whether it belongs to the production area.

[0029] In some embodiments, the number of the traceability models is consistent with the number of production area labels; the traceability models include at least two, namely, the production area one traceability model and the production area two traceability model;

[0030] Optionally, the input is processed in the traceability model of different production areas to obtain a classification result of whether it belongs to the production area, including: inputting the metallomics data and / or plant metabolomics data into the traceability model of production area one, and outputting a yes or no result; if a yes result is output, ending the operation, and obtaining a result that the sample to be tested belongs to production area one; if a no result is output, obtaining a result that the sample to be tested does not belong to production area one, and inputting the metallomics data and / or plant metabolomics data into the traceability model of production area two, and outputting a yes or no result; if a yes result is output, ending the operation, and obtaining a result that the sample to be tested belongs to production area two; if a no result is output, obtaining a result that the sample to be tested does not belong to production area two.

[0031] The fourth aspect of the present application discloses a computer device, which includes: a memory and a processor; the memory is used to store a computer program; and the processor executes the computer program to implement the steps of the above method.

[0032] A fifth aspect of the present application discloses a computer-readable storage medium having a computer program stored thereon, which implements the steps of the above method when the computer program is executed by a processor.

[0033] A sixth aspect of the present application discloses a computer program product, including a computer program, which implements the steps of the above method when executed by a processor.

[0034] This application has the following beneficial effects:

[0035] 1. In the process of constructing a model in the field of traditional Chinese medicine, this application replicates or amplifies minority data points (such as samples of Radix Peucedanum with a specific origin and a specific ingredient content) so that the number of samples of each category in the training data tends to be balanced. Accordingly, the algorithm will no longer easily ignore those important but scarce categories during the training process, so that it can learn the characteristics and information in the data more comprehensively and improve the model's recognition ability for minority samples. Of course, oversampling technology also has a certain risk of overfitting. In the field of traditional Chinese medicine detection, this may cause the model to overfit the minority samples in the training data, and perform poorly in actual applications. In order to reduce this risk, reasonable parameter settings and cross-validation can be adopted. For example, the model performance can be optimized by adjusting the oversampling ratio, selecting appropriate machine learning algorithms and parameters, and other methods. In general, applying oversampling technology to the field of traditional Chinese medicine detection can significantly offset the negative effects of unbalanced learning, improve the recognition ability of machine learning models for minority samples, and provide more accurate and reliable solutions for practical application scenarios such as traditional Chinese medicine quality control and efficacy evaluation;

[0036] 2. This application constructs and applies an aggregate model to more accurately and efficiently predict the origin of Peucedanum chinense. The aggregate model has the following advantages: (1) Improving prediction accuracy. By integrating the results of multiple prediction models, the aggregate model can combine the advantages of different models and reduce the prediction error of a single model, thereby improving the prediction accuracy. In the prediction of the origin and authenticity of Peucedanum chinense, different models may make predictions based on different characteristics (such as climate, soil, ecological environment, etc.). The aggregate model can comprehensively consider these characteristics and give more accurate prediction results. (2) Enhancing the generalization ability of the model. By combining multiple models, the aggregate model can reduce the risk of overfitting of a single model to specific data and improve the generalization ability of the model. This means that the aggregate model can give more reliable prediction results when faced with new and unknown Peucedanum chinense origin data. This is of great significance for the cultivation, procurement and quality control of Chinese medicinal materials. (3) Making full use of multiple data sources and information. The origin and authenticity of Chinese medicinal materials are affected by many factors, including climate, soil, ecological environment, cultivation technology, etc. The aggregation model can integrate information from different data sources, such as meteorological data, soil data, remote sensing images, etc., as well as expert judgment based on traditional knowledge and experience, so as to more comprehensively evaluate the origin and authenticity of Peucedanum chinense. (4) Improve prediction efficiency. The aggregation model can significantly improve the prediction efficiency by processing multiple prediction models in parallel. This is of great significance for the rapid response and decision-making of the Chinese medicinal materials industry. For example, in the process of Chinese medicinal materials planting, timely and accurate prediction of the origin and authenticity of Peucedanum chinense can help guide growers to choose appropriate planting areas and cultivation techniques, and improve the yield and quality of Chinese medicinal materials. (5) Adaptability and flexibility. The aggregation model has high adaptability and flexibility. It can select different prediction models for combination according to actual needs and data characteristics, and adjust the combination method and weight. This enables the aggregation model to adapt to different application scenarios and data distributions, and provide more personalized solutions. For example, when predicting the origin and authenticity of Peucedanum chinense for a specific region or a specific variety, the aggregation model can select a more suitable prediction model for combination according to the climate, soil and other characteristics of the region.

[0037] 3. This application innovatively proposes a combined identification model of element fingerprint substances and secondary metabolite profile substances for the identification of the geographical origin of Chinese medicine with Peucedanum chinense as a model. Based on metallomics and plant metabolomics, the element fingerprint is integrated with the chemical component profile. The results show that compared with chemometrics, machine learning modeling with the help of data oversampling technology can achieve high-precision identification of production areas in Anhui, Zhejiang, Guizhou and Chongqing. Here, the innovative use of aggregation models further integrates the optimal models of elements and chemical components to improve classification accuracy. In addition, variable importance gives a professional explanation of the models of elements and chemical components. In short, for the first time, based on the perspective of machine learning fusion of element fingerprints and secondary metabolite profile factors, a comprehensive insight into the scientific connotation of the production area of ​​Peucedanum chinense was obtained, which provides a scientific basis for scientific and reasonable quality supervision and clinical application. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative work.

[0039] Figure 1 is a schematic diagram of a method flow chart provided by the first aspect of an embodiment of the present invention;

[0040] Figure 2 is a schematic diagram of a method flow chart provided by the second aspect of an embodiment of the present invention;

[0041] Figure 3 is a schematic diagram of a method flow chart provided by the third aspect of an embodiment of the present invention;

[0042] Figure 4 It is a schematic diagram of a system for constructing a Chinese medicine origin tracing model integrating element fingerprints and secondary metabolite profiles provided in an embodiment of the present invention;

[0043] Figure 5 It is a schematic diagram of a Chinese medicine origin tracing system integrating element fingerprints and secondary metabolite profiles provided by an embodiment of the present invention;

[0044] Figure 6 It is a schematic diagram of a Chinese medicine origin tracing system integrating element fingerprints and secondary metabolite profiles provided by another embodiment of the present invention;

[0045] Figure 7 is a schematic diagram of a computer device provided by an embodiment of the present invention;

[0046] Figure 8 is a schematic diagram of the architecture of an exemplary computing device provided by an embodiment of the present invention;

[0047] Fig. 9 is a schematic diagram of a storage medium provided by an embodiment of the present invention;

[0048] Fig.10 is a schematic diagram of the chemometric analysis results of the element fingerprint profile provided by an embodiment of the present invention; wherein, Fig.10 A is the PCA result analysis between different production areas. Fig.10 B is the PLSDA analysis result, Fig.10 C is the cluster heat map analysis result;

[0049] Fig.11 is a schematic diagram of the chemometric analysis results of the metabolomics profile provided in an embodiment of the present invention; wherein, Fig.11 A is the PCA result analysis between different production areas. Fig.11 B is the PLSDA analysis result, Fig.11 C is the cluster heat map analysis result;

[0050] Fig.12 : is a schematic diagram of ROC curves of different production areas established by integrating element fingerprints with five machine learning algorithms provided by an embodiment of the present invention; wherein, Fig.12 A is Anhui, Fig.12 B is Guizhou, Fig.12 C stands for Zhejiang, Fig.12 D stands for Chongqing;

[0051] Fig.13 is a schematic diagram of a correlation analysis diagram of secondary metabolites before and after correlation removal provided by an embodiment of the present invention; wherein, Fig.13 A is before removal, Fig.13 B is after removal;

[0052] Fig.14 : is a schematic diagram of ROC curves of different production areas established by integrating five machine learning algorithms for secondary metabolites provided in an embodiment of the present invention; wherein, Fig.14 A is Anhui, Fig.14 B is Guizhou, Fig.14 C stands for Zhejiang, Fig.14 D stands for Chongqing;

[0053] Fig.15 Schematic diagram of ROC curves of different production areas established by the aggregation model of metallomics fusion plant metabolomics provided in an embodiment of the present invention; wherein, Fig.15 A is Guizhou, Fig.15 B is Chongqing, Fig.15 C is Anhui, Fig.15 D stands for Zhejiang;

[0054] Fig.16 are the top fifteen representative components or elements with outstanding importance of Anhui production area variables provided by the embodiment of the present invention; among them, Fig.16 A and B are the importance distribution of the element model. Fig.16 C and D are the importance distribution of chemical components;

[0055] Fig.17 are the top fifteen representative components or elements with outstanding importance of Guizhou production area variables provided by the embodiment of the present invention; among them, Fig.17 A and B are the importance distribution of the element model. Fig.17 C and D are the importance distribution of chemical components;

[0056] Fig.18 are the top fifteen representative components or elements with outstanding importance of Zhejiang production area variables provided by the embodiment of the present invention; among them, Fig.18 A and B are the importance distribution of the element model. Fig.18 C and D are the importance distribution of chemical components;

[0057] Fig.19 are the top fifteen representative components or elements with outstanding importance of Chongqing production area variables provided by the embodiment of the present invention; among them, Fig.19 A and B are the importance distribution of the element model. Fig.19 C and D show the importance distribution of chemical components. DETAILED DESCRIPTION

[0058] In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0059] In some of the processes described in the specification and claims of the present invention and the above-mentioned figures, multiple operations that appear in a specific order are included, but it should be clearly understood that these operations may not be executed in the order in which they appear in this article or executed in parallel. The serial numbers of the operations, such as 101, 102, etc., are only used to distinguish different operations, and the serial numbers themselves do not represent any execution order. In addition, these processes may include more or fewer operations, and these operations may be executed in sequence or in parallel. It should be noted that the descriptions of "first", "second", etc. in this article are used to distinguish different messages, devices, modules, etc., do not represent the order of precedence, and do not limit the "first" and "second" to be different types.

[0060] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0061] Figure 1 The present invention provides a schematic flow chart of a method for constructing a Chinese medicine origin tracing model integrating element fingerprints and secondary metabolite profiles. Specifically, the method comprises the following steps:

[0062] 101: Obtain the CPS response value of each element in the training set samples from different production areas and the first production area label corresponding to the sample;

[0063] In some embodiments, the first production area labels include at least two, production area 1 and production area 2;

[0064] In some embodiments, the training set samples include: Radix Peucedanum, Radix Peucedanum purpurogenum, Radix Saposhnikoviae, etc.; when the training set samples are Radix Peucedanum, the first production area label or the second production area label includes any two or more of the following: Anhui, Zhejiang, Chongqing, Guizhou;

[0065] In some embodiments, the element fingerprint material is the metallomics including the element fingerprint, and the secondary metabolite profile material is the plant metabolomics including the secondary metabolite profile. Each element includes a major element and a trace element; the trace element includes a metal element and a non-metal element.

[0066] 102: Input the CPS response value into the first machine learning model to obtain a prediction classification result, compare it with the first production area label, and optimize the model according to the comparison result to obtain a metallomics model;

[0067] In some embodiments, between 101 and 102, the CPS response value is also included in preprocessing, and the preprocessing includes: 1011, preliminary standardization processing of dividing the CPS response value by the sampling amount; after 1011, the preprocessing also includes: 1012, normalizing the CPS response values ​​of major elements and trace elements in the training set samples from different production areas; performing dimensionality reduction processing on the normalized CPS response value; the dimensionality reduction processing methods include PCA, OPLS-DA and cluster heat map analysis.

[0068] 103: Obtaining mass spectrometry data of secondary metabolites of training set samples from different production areas and second production area labels corresponding to the samples; the second production area label is consistent with the first production area label;

[0069] 104: inputting the mass spectrometry data into a second machine learning model to obtain a prediction classification result, comparing it with the second production area label, optimizing the model according to the comparison result, and obtaining a plant metabolomics model;

[0070] In some embodiments, between 103 and 104, the step further includes performing a cleaning process on the mass spectrum data to reduce redundancy and highly correlated features, thereby obtaining cleaned mass spectrum data;

[0071] Optionally, the cleaning process includes: 1031, taking the compound information with a response value greater than a first threshold as the basic data set, removing the molecular ion peaks close to the mass spectrum baseline and with high accuracy; 1032, performing correlation removal processing on the molecular ion peaks close to the mass spectrum baseline and with high accuracy that have been removed, to obtain secondary metabolite mass spectrum data after removal; Optionally, the first threshold is 1×10 5 The first threshold is obtained through training of training set samples, and may be a specific threshold or an interval range, and the specific form is not specifically limited in this embodiment.

[0072] 105: Select the optimal model of the metallomics model and plant metabolomics model in the same production area, aggregate the two optimal models to obtain the aggregated model of the same production area as the traceability model of the production area.

[0073] In a more specific embodiment, the 105 includes: selecting the optimal model in the metallomics model of production area one, selecting the optimal model in the plant metabolomics model of production area one, and aggregating the two optimal models to obtain the traceability model of production area one; selecting the optimal model in the metallomics model of production area two, selecting the optimal model in the metabolomics model of production area two, and aggregating the two optimal models to obtain the traceability model of production area two.

[0074] In a more specific embodiment, the aggregation method of the aggregation model includes: bootstrap aggregation, boosting, stacking, mixing, etc.

[0075] In some embodiments, when the different production areas include 3, the production area label also includes production area 3; the optimal model in the metallomics model of production area 3 is selected, the optimal model in the metabolomics model of production area 3 is selected, and the two optimal models are aggregated to obtain the traceability model of production area 3; optionally, when the different production areas include 4, the production area label also includes production area 4; the optimal model in the metallomics model of production area 4 is selected, the optimal model in the metabolomics model of production area 4 is selected, and the two optimal models are aggregated to obtain the traceability model of production area 4;

[0076] In some embodiments, the methods used by the first machine learning model and / or the second machine learning model include: RF, KNN, Rpa, SVM, XGB.

[0077] In some embodiments, when the training set sample is a Radix Peucedanum medicinal material, the first production area label or the second production area label includes any two or more of the following: Anhui, Zhejiang, Chongqing, and Guizhou; in the first machine learning model, the RF algorithm used in the Anhui production area has high accuracy, the KNN algorithm used in the Zhejiang production area has high accuracy, the XGB algorithm used in the Chongqing production area has high accuracy, the SVM algorithm used in the Guizhou production area has high accuracy, and the XGB algorithm used in the authentic production area has high accuracy. In the second machine learning model, the KNN algorithm used in the Anhui production area has high accuracy, the XGB algorithm used in the Zhejiang production area has high accuracy, the SVM algorithm used in the Chongqing production area has high accuracy, the SVM algorithm used in the Guizhou production area has high accuracy, and the KNN algorithm used in the authentic production area has high accuracy. Specifically, during aggregation: the model of Anhui production area using RF algorithm in the metallomics model and the model of Anhui production area using KNN algorithm in the metabolomics model were selected for aggregation to obtain the aggregation model of Anhui production area; the model of Zhejiang production area using KNN algorithm in the metallomics model and the model of Zhejiang production area using XGB algorithm in the metabolomics model were selected for aggregation to obtain the aggregation model of Zhejiang production area; the model of Chongqing production area using XGB algorithm in the metallomics model and the model of Chongqing production area using SVM algorithm in the metabolomics model were selected for aggregation to obtain the aggregation model of Chongqing production area; the model of Guizhou production area using SVM algorithm in the metallomics model and the model of Guizhou production area using SVM algorithm in the metabolomics model were selected for aggregation to obtain the aggregation model of Guizhou production area; the model of authentic production area using XGB algorithm in the metallomics model and the model of authentic production area using KNN algorithm in the metabolomics model were selected for aggregation to obtain the aggregation model of authentic production area.

[0078] In some embodiments, the present application discloses a method for tracing the origin of traditional Chinese medicine by integrating element fingerprints and secondary metabolite profiles, such as Figure 2 As shown, the method includes:

[0079] 201, obtain metallomics data and plant metabolomics data of the samples to be tested;

[0080] 202, inputting the metallomics data and / or plant metabolomics data into the traceability model of different production areas for processing, and obtaining a classification result of whether the data belongs to the production area;

[0081] Optionally, the metallomics data includes CPS response values ​​of any one or several of the following characteristic elements: Mn element, S element, Mg element, B element; optionally, the plant metabolomics data includes mass spectrometry data of any one or several of the following chemical component peak information: peak 16696, peak 13779, peak 19774, peak 17630.

[0082] In some embodiments, the present application discloses a method for tracing the origin of traditional Chinese medicine by integrating element fingerprints and secondary metabolite profiles, such as Figure 3As shown, the method includes:

[0083] 301, obtaining CPS response values ​​and mass spectrum data of target peak information of target characteristic elements of the sample to be tested, wherein the target characteristic elements include any one or more of the following: Mn element, S element, Mg element, and B element; the target peak information includes any one or more of the following: peak 16696, peak 13779, peak 19774, and peak 17630;

[0084] 302, inputting the CPS response value of the target characteristic element and the mass spectrum data of the target peak information into the traceability model of different production areas for processing, and obtaining a classification result of whether it belongs to the production area.

[0085] In some embodiments, the number of the traceability models is consistent with the number of production area labels; the traceability models include at least two, namely, the production area one traceability model and the production area two traceability model;

[0086] Optionally, the input is processed in the traceability model of different production areas to obtain a classification result of whether it belongs to the production area, including: inputting the metallomics data and / or plant metabolomics data into the traceability model of production area one, and outputting a yes or no result; if a yes result is output, ending the operation, and obtaining a result that the sample to be tested belongs to production area one; if a no result is output, obtaining a result that the sample to be tested does not belong to production area one, and inputting the metallomics data and / or plant metabolomics data into the traceability model of production area two, and outputting a yes or no result; if a yes result is output, ending the operation, and obtaining a result that the sample to be tested belongs to production area two; if a no result is output, obtaining a result that the sample to be tested does not belong to production area two.

[0087] In some embodiments, if the result is that the sample to be tested does not belong to production area two, the metallomics data and / or plant metabolomics data can be input into the traceability model of production area three, and a yes or no result is output; if a yes result is output, the operation is terminated, and a result is obtained that the sample to be tested belongs to production area three; if a no result is output, a result is obtained that the sample to be tested does not belong to production area three, and the metallomics data and / or plant metabolomics data are input into the traceability model of production area four, and a yes or no result is output; if a yes result is output, the operation is terminated, and a result is obtained that the sample to be tested belongs to production area four; if a no result is output, a result is obtained that the sample to be tested does not belong to production area four, and so on. In this embodiment, taking the previous Hu as an example, production area one, production area two, production area three, and production area four are any sorted combination of Zhejiang, Anhui, Guizhou, and Chongqing.

[0088] Figure 7 is a schematic diagram of a computer device provided by an embodiment of the present invention, such as Figure 7As shown, the device may include: one or more processors, and one or more memories; wherein the memories store computer-readable codes, and when the computer-readable codes are run by the one or more processors, the method described above may be executed.

[0089] The processor in this embodiment can be an integrated circuit chip with signal processing capabilities. The above processor can be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The disclosed methods, operations and logic block diagrams in the embodiments of the present disclosure can be implemented or executed. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc., which can be an X86 architecture or an ARM architecture.

[0090] In general, various example embodiments of the present disclosure may be implemented in hardware or dedicated circuits, software, firmware, logic, or any combination thereof. Certain aspects may be implemented in hardware, while other aspects may be implemented in firmware or software that may be executed by a controller, microprocessor, or other computing device. When various aspects of the disclosed embodiments are illustrated or described as block diagrams, flow charts, or using some other graphical representation, it will be understood that the blocks, devices, systems, techniques, or methods described herein may be implemented in hardware, software, firmware, dedicated circuits or logic, general purpose hardware or controllers or other computing devices, or some combination thereof as non-limiting examples.

[0091] For example, the method or device according to the embodiment of the present disclosure may also be implemented by Figure 8 The architecture of the computing device 3000 shown in FIG. Figure 8 As shown, the computing device 3000 may include a bus 3010, one or more CPUs 3020, a read-only memory (ROM) 3030, a random access memory (RAM) 3040, a communication port 3050 connected to a network, an input / output component 3060, a hard disk 3070, etc. The storage device in the computing device 3000, such as ROM 3030 or hard disk 3070, may store various data or files used for processing and / or communication of the method provided by the present disclosure and program instructions executed by the CPU. The computing device 3000 may also include a user interface 3080. Of course, Figure 8 The architecture shown is only exemplary and can be omitted according to actual needs when implementing different devices. Figure 8 One or more components of a computing device are shown.

[0092] The embodiment of the present invention also provides a computer-readable storage medium, such as Fig. 9As shown, it is a schematic diagram of a storage medium provided by an embodiment of the present invention, and a computer readable instruction 4010 is stored on the computer storage medium 4020. When the computer readable instruction 4010 is executed by the processor, the method according to the embodiment of the present disclosure described with reference to the above figures can be executed. The computer readable storage medium in the embodiment of the present disclosure can be a volatile memory or a non-volatile memory, or can include both volatile and non-volatile memories. The non-volatile memory can be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM) or a flash memory. The volatile memory can be a random access memory (RAM), which is used as an external cache. By way of example and not limitation, many forms of RAM are available, such as static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDRSDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous link dynamic random access memory (SLDRAM), and direct memory bus random access memory (DR RAM). It should be noted that the memory of the methods described herein is intended to include, but is not limited to, these and any other suitable types of memory. It should be noted that the memory of the methods described herein is intended to include, but is not limited to, these and any other suitable types of memory.

[0093] The embodiments of the present disclosure further provide a computer program product or system, including a computer program, which implements the steps of the above method when executed by a processor.

[0094] In some embodiments, this embodiment also discloses a system for constructing a Chinese medicine origin tracing model that integrates element fingerprints and secondary metabolite profiles, such as Figure 4 As shown, the system comprises:

[0095] The first acquisition module 401 is used or configured to obtain the CPS response value of each element in the training set samples from different production areas and the first production area label corresponding to the sample;

[0096] A metallomics model building module 402 is used or configured to input the CPS response value into a first machine learning model to obtain a prediction classification result, compare it with the first production area label, and optimize the model according to the comparison result to obtain a metallomics model;

[0097] The second acquisition module 403 is used or configured to acquire mass spectrometry data of secondary metabolites of training set samples from different production areas and second production area labels corresponding to the samples; the second production area labels are consistent with the first production area labels;

[0098] A plant metabolomics model building module 404 is used or configured to input the mass spectrometry data into a second machine learning model to obtain a prediction classification result, compare it with the second production area label, optimize the model according to the comparison result, and obtain a plant metabolomics model;

[0099] The model aggregation module 405 is used for or configured to select the optimal model of the metallomics model and the plant metabolomics model in the same production area, aggregate the two optimal models to obtain an aggregated model of the same production area as the traceability model of the production area.

[0100] In some embodiments, the system further includes: a preprocessing module located between 401 and 402, used for or configured to preprocess the CPS response value, the preprocessing including: 1011, preliminary standardization of the CPS response value by dividing the sampling amount; optionally, after 1011, the preprocessing further includes: 1012, normalizing the CPS response values ​​of constant elements and trace elements in training set samples from different production areas; performing dimensionality reduction processing on the normalized CPS response value; the dimensionality reduction processing methods include PCA, OPLS-DA and cluster heat map analysis.

[0101] In some embodiments, the system further includes: a data cleaning processing module located between 403 and 404, used for or configured to perform cleaning processing on the mass spectrometry data to reduce redundancy and highly correlated features, to obtain cleaned mass spectrometry data; optionally, the cleaning processing includes: 1031, taking compound information with a response value greater than a first threshold as a basic data set, removing accurate molecular ion peaks close to the mass spectrometry baseline; 1032, performing correlation removal processing on the accurately removed molecular ion peaks close to the mass spectrometry baseline, to obtain removed secondary metabolite mass spectrometry data.

[0102] In some embodiments, the system further includes: an oversampling processing module located between 401 and 402, which is used or configured to replicate or amplify minority class data points to balance the number of samples of each category in the training data; the minority class data points include any one or more of the following: specific origin, special ingredient content.

[0103] In some embodiments, this embodiment also discloses a Chinese medicine origin tracing system integrating element fingerprints and secondary metabolite profiles, such as Figure 5 As shown, the system comprises:

[0104] A third acquisition module 501 is used or configured to acquire metallomics data and plant metabolomics data of the sample to be tested;

[0105] The first production area classification output module 502 is used for or configured to input the metallomics data and / or plant metabolomics data into the traceability model of different production areas for processing to obtain a classification result of whether it belongs to the production area.

[0106] In some embodiments, 502 includes: the number of the traceability models is consistent with the number of production area labels; the traceability models include at least 2, namely, production area one traceability model and production area two traceability model; the input is processed in the traceability models of different production areas to obtain a classification result of whether it belongs to the production area, including: inputting the metallomics data and / or plant metabolomics data into the production area one traceability model, and outputting a yes or no result; if the output is a yes result, the operation is terminated, and the result that the sample to be tested belongs to production area one is obtained; if the output is a no result, the result that the sample to be tested does not belong to production area one is obtained, and the metallomics data and / or plant metabolomics data are input into the production area two traceability model, and a yes or no result is output; if the output is a yes result, the operation is terminated, and the result that the sample to be tested belongs to production area two is obtained; if the output is a no result, the result that the sample to be tested does not belong to production area two is obtained.

[0107] In some embodiments, this embodiment also discloses another Chinese medicine origin tracing system integrating element fingerprints and secondary metabolite profiles, such as Figure 6 As shown, the system comprises:

[0108] The fourth acquisition module 601 is used or configured to obtain the mass spectrum data of the CPS response value and target peak information of the target characteristic element of the sample to be tested, wherein the target characteristic element includes any one or more of the following: Mn element, S element, Mg element, and B element; the target peak information includes any one or more of the following: peak 16696, peak 13779, peak 19774, and peak 17630;

[0109] The second production area classification output module 602 is used for or configured to input the CPS response value of the target characteristic element and the mass spectrum data of the target peak information into the traceability model of different production areas for processing to obtain a classification result of whether it belongs to the production area.

[0110] In some embodiments, the number of traceability models is consistent with the number of production area labels; the traceability models include at least 2, namely, production area one traceability model and production area two traceability model; the input is processed in the traceability models of different production areas to obtain a classification result of whether it belongs to the production area, including: inputting the metallomics data and / or plant metabolomics data into the production area one traceability model, and outputting a yes or no result; if the output is a yes result, the operation is terminated, and the result that the sample to be tested belongs to production area one is obtained; if the output is a no result, the result that the sample to be tested does not belong to production area one is obtained, and the metallomics data and / or plant metabolomics data are input into the production area two traceability model, and a yes or no result is output; if the output is a yes result, the operation is terminated, and the result that the sample to be tested belongs to production area two is obtained; if the output is a no result, the result that the sample to be tested does not belong to production area two is obtained. Specific embodiment:

[0112] Peucedanum chinense is a traditional Chinese medicinal material with a long history. It has a prominent position in traditional Chinese medicine and is one of the key medicinal flavors in ancient classic prescriptions. Peucedanum chinense has a history of more than 1,500 years of medicinal use in my country, and was first recorded in "Famous Doctors' Records". It has been recorded in all dynasties' herbal books and has been included in the "Chinese Pharmacopoeia" since 1963. After 2005, white flower Peucedanum chinense became the main variety, and purple flower Peucedanum chinense was included separately in 2010. In the current 2020 edition of the "Chinese Pharmacopoeia", Peucedanum chinense refers to the dried root of white flower Peucedanum chinense. Peucedanum chinense is distributed in many provinces and regions in my country, and is most famous in the areas around Tianmu Mountain, such as Anhui, Zhejiang, and Jiangxi. In particular, "Faithful Peucedanum chinense" is recognized as an authentic medicinal material. With the growth of market demand, Peucedanum chinense has gradually developed from wild to simulated wild, woodland and field planting. The main planting areas have expanded to Ningguo, Anhui, Hangzhou, Zhejiang, Chongqing, Sichuan, Guizhou and other places.

[0113] This study uses Peucedanum chinense as a model Chinese medicine. Aiming at the problem of origin traceability and quality identification, it innovatively integrates non-targeted and targeted metallomics, plant metabolomics and machine learning technologies to build a set of efficient and accurate analysis strategies. By comprehensively analyzing the elemental fingerprint and secondary metabolite profile of Peucedanum chinense, combined with the construction of intelligent models, it is possible to quickly and accurately identify the origin of Peucedanum chinense. This study can provide new ideas, methods and tools for the quality evaluation of Peucedanum chinense, and set a benchmark for the improvement of the overall quality evaluation system of traditional Chinese medicine and scientific supervision.

[0114] 1. Materials and Methods

[0115] 1.1 Materials and reagents

[0116] 1.1.1 Instruments: Inductively coupled plasma mass spectrometer (Shimadzu Corporation, Japan, model 2030LF); microwave digester (CEM Corporation, USA, model MARS6); electronic balance (METTLER TOLEDO, model XSE205DU); ultrapure water system (Millipore Corporation, USA, model Milli-QAdvantage A10); acid-removing heater (Beijing Donghang Technology Instrument Co., Ltd., model BHW-C); ultra-high performance liquid chromatography (Thermo Scientific, model Vanquish); Orbitrap Fusion Lumos Tribrid three-in-one mass spectrometer (Thermo Scientific), equipped with Xcalibur3.2 software processing system and Compound Discover 2.0 software (Thermo Scientific).

[0117] 1.1.2 Reagents: Chromatographic grade nitric acid (Sigma, USA); Al (aluminum), Ag (silver), As (arsenic), Ba (barium), B (boron), Be (beryllium), Ca (calcium), Cu (copper), Cd (cadmium), Co (cobalt), Cr (chromium), Fe (iron), Hg (mercury), K (potassium), Mg (magnesium), Mn (manganese), Mo (molybdenum), Na (sodium), Ni (nickel), Pb (lead), Rb (rubidium), Sr (strontium), Sb (antimony), Se (selenium), Sn (tin), Ti (titanium), Tl (thallium), V (vanadium), Zn (zinc) single element standard solutions were purchased from the National Institute of Metrology, China; the tuning solution was a mixed standard solution of Li, Y, Ce, Tl, and Co (1 μg / L); the internal standard solution was a standard solution containing Ge 100 μg / mL mixed internal standard solution (Agilent, USA, batch number 5188-6525); chromatographic methanol (DiKMA, mass spectrometry grade), chromatographic acetonitrile (DiKMA, mass spectrometry grade), mass spectrometry grade formic acid (Thermo Fisher Scientific, mass spectrometry grade, batch number A117-50), and deionized water were purified by Milli-Q.

[0118] 1.1.3 Samples: A total of 73 batches of white flower peucedanum were collected in this study, mainly from major production areas, medicinal material markets, medicinal material companies, pharmacies, etc. The main production areas are from Zhejiang, Anhui, Chongqing, Guizhou and other production areas. All samples were identified by Chief Pharmacist Jin Hongyu of China Food and Drug Inspection Institute as the dried roots of Peucedanum peucedanum Dunn, a plant of the Umbelliferae family. The sample specimens were stored in the laboratory of the Institute of Traditional Chinese Medicine and Ethnomedicine Inspection of China Food and Drug Inspection Institute under cool conditions.

[0119] 1.2 Metallomics analysis method: After the sample is crushed, it is passed through a No. 4 sieve. 0.5 g of the sample to be tested is accurately weighed and placed in a microwave digestion tank. 8.0 mL of nitric acid is added. The device is installed according to the operating procedures and digested (heating to 120°C for 3 minutes, heating to 150°C for 2 minutes, heating to 200°C for 2 minutes, and holding for 12 minutes). After the digestion is completed, it is cooled to below 60°C, and the digestion tank is placed on a heating plate at 100°C to drive the acid until no brown-yellow acid mist is emitted. The digestion tank is taken out and cooled. The digestion solution is transferred to a 50 mL volumetric flask, and the digestion tank is washed 3 times with a small amount of water. The washing solution is combined in the volumetric flask, and the volume is adjusted to the scale with water. Shake well to obtain the test solution. The reagent blank solution is prepared in the same way.

[0120] This study used the collision reaction cell mode. During the ICP-MS operation, the high-frequency power was 1.20 kW, the carrier gas (high-purity argon) flow rate was 0.70 L / min, the plasma gas flow rate was 9.0 L / min, the auxiliary gas flow rate was 1.1 L / min, the peristaltic pump speed was 0.3 r / s, the sampling depth was 7 mm, the cell gas was 6.0 mL / min, the cell voltage was -21 V, the energy filter was 7.0 V, and the nebulizer temperature was 5 °C. The elements and isotopes analyzed semi-quantitatively included 107 Ag, 27 Al, 75 As, 197 Au, 11 B. 138 8. 9 Be, 79 Br, 44 Ca, 114 Cd, 140 Ce, 35 Cl, 59 Co. 52 Cr, 133 Cs, 63 Cu, 164 Dy, 166 2. 153 Eu, 56 Fe, 69 Ga, 158 Gd, 180 Hf 202 Hg, 165 Ho, 127 I. 193 Ir, 39 K. 139 La, 7 Li, 24 Mg, 55 Mn, 95 Mo, 23 Na, 93 Nb,142 Nd, 60 You, 192 Os, 31 P. 208 Pb, 108 Pd, 141 Pr, 195 Pt, 85 Rb, 187 Re, 102 Such as 34 S. 121 Sb, 78 Se, 28 4. 152 Sm, 118 Sn, 88 Sr. 181 She 130 Te, 232 Th, 47 Ti, 205 Tl, 169 Tm, 238 U. 51 V. 184 W. 89 Y. 174 Yb, 66 Zn, 90 Zr. Accurately measure an appropriate amount of internal standard solution, place it in a volumetric flask, and dilute it to the scale with 5% nitric acid solution. 72 Ge, 6 Li, 45 Sc, 115 In, 209 Bi, 175 Lu, 103 Rh, 159 The Tb concentration was 500 ng / mL, and the internal standard solution was obtained. Combining the principle of determination stability and similar mass number, a suitable internal standard was selected. The sample tube was inserted into the standard solution and sample solution in turn for determination. The CPS response value of each element was obtained by internal standard correction and subtraction of blank background.

[0121] 1.3 Metabolite profile analysis method: Take about 0.5 g of Peucedanum chinense powder (passed through No. 3 sieve), weigh it accurately, put it in a stoppered conical flask, accurately add 25 ml of methanol, stopper it, weigh it, ultrasonically treat it (power 250 W, frequency 33 kHz) for 30 minutes, cool it, shake it well, and filter it to obtain the sample solution.

[0122] Chromatographic conditions: ACCQUITY UHPLC HSS T3 (2.1×100 mm, 1.8 μm) column, acetonitrile (A)-0.1% formic acid (B) system as mobile phase, gradient elution program as follows: 0-3 min, 90%-70% B; 3-7 min, 70%-50% B; 7-20 min, 50%-40% B; 20-30 min, 40%-10% B; 30-32 min, 10% B; 32-35 min, 10%-90%. Flow rate 0.3 mL / min; column temperature 30°C; injection volume 2 μL.

[0123] Mass spectrometry conditions: Heated electrospray ion source (H-ESI) detection mode, positive ion mode detection, nebulizer gas is high-purity nitrogen (Ar), collision gas is high-purity helium (He). Sheath gas flow rate is 45L / min, auxiliary gas flow rate is 15L / min, spray voltage is 3.5kV in positive ion mode, capillary temperature is 350℃, auxiliary temperature is 320℃; primary scanning mode is MS OT, mass range is m / z70~1000, resolution is 120000, injection time is 15ms; secondary mass spectrometry scanning detector is Orbitrap (dd-MS2 OT HCD), resolution is 30000; collision energy (CE) is 15, 30, 50eV.

[0124] 1.4 General data processing method: The multi-element CPS values ​​obtained by semi-quantitative collection were initially standardized by dividing the CPS by the sampling volume. According to the distribution of different production areas, the R language chemometrics package was used to first perform normalization, and then PCA, OPLS-DA and cluster heat map analysis were performed to obtain general chemometric characteristics through data dimensionality reduction.

[0125] The plant metabolomics processing method is as follows: the collected raw files are input into QI software (Waters Company) for pre-processing, including peak extraction and peak matching, to preliminarily obtain the ion response intensity in each sample. The R language chemometrics package is used to first perform normalization processing, and then PCA, OPLS-DA and cluster heat map analysis are performed to obtain general chemometric features through data dimensionality reduction.

[0126] 1.5 Machine Learning Modeling Methods

[0127] 1.5.1 Dataset division: In the study based on the semi-quantitative element CPS response data and comprehensive metabolomics secondary metabolite data of 73 batches of Radix Peucedanum, this study carefully planned the data utilization strategy. This study explicitly used non-targeted metal group data, metabolomics data, and the combination of the two as the cornerstones for building machine learning models. The amount of secondary metabolite data is huge, and correlation removal is used before data set division to reduce metabolite-related features and improve computational efficiency. In order to ensure the generalization ability of the model and reduce the risk of overfitting, this study adopted a scientific data division method, that is, the data was randomly and evenly divided into a training set (70%) and a test set (30%). The training set was fully utilized to train and optimize the model, while the test set was used as an independent validation set to objectively evaluate the model's key performance indicators such as accuracy and error. This optimization strategy not only retains the original intention, but also further improves the rigor and reliability of the study through clear data division and scientific evaluation methods.

[0128] 1.5.2 Selection of machine learning methods: Combined with the profound previous research foundation of our research group, this study will make full use of the powerful functions of R and Python languages, focus on the binary classification algorithm framework, adopt data oversampling strategy, and deal with data imbalance. Then, six efficient and complementary machine learning algorithms, kNN (K nearest neighbor), RF (random forest), rpart (recursive partition tree), Stacking (stacked ensemble), SVM (support vector machine) and XGBoost (extreme gradient boosting) under the mlr3 framework based on R were carefully selected and integrated. The aggregation model was used to integrate the optimal model of metallomics and plant metabolomics. The core goal of this study was to optimize and construct an advanced model for accurately identifying the main production areas and authenticity of Radix Peucedanum. Through the in-depth analysis and application of this model, this study aims to further explore and reveal the complex scientific laws behind the quality of Radix Peucedanum, and provide strong data support and scientific basis for the quality evaluation, origin traceability and authenticity certification of Chinese herbal medicines. This optimization scheme emphasizes the diversity and complementarity of algorithm selection, as well as the scientificity and foresight of model construction on the basis of retaining the original content.

[0129] 1.5.3 Construction and evaluation of classification model: In the process of constructing and evaluating the classification model, this study adopted a 5-fold cross-validation strategy, which is an effective model evaluation method that can more comprehensively test the generalization ability of the model. This study selected accuracy (ACC) and area under the curve (AUC) as core evaluation indicators, which jointly measure the performance of the model. In particular, AUC is given a primary position in the process of model parameter optimization because it can more comprehensively reflect the classification ability of the model under different thresholds. In order to ensure the reliability and stability of model evaluation, this study further strengthened the experimental design and repeated the modeling process for 10 cycles. This measure not only helps to reduce the impact of random errors on the results, but also obtains more robust model performance evaluation results through multiple iterations. In each cycle, this study will strictly follow the 5-fold cross-validation process to ensure full use of data and fairness of model evaluation.

[0130] 1.5.4 Variable Importance Analysis: In this study, the output of the model is explained by assigning importance values ​​to the features. The fastshap package is used to calculate variable importance, and the shapviz package is used to visualize variable importance. This study introduces SHAP into the interpretation of machine learning models to calculate the contribution of each feature to the model prediction results. Through the SHAP value, we can intuitively see the impact of the feature on the model prediction results, and explain the prediction results of a single data point, which helps us understand why the model makes a certain prediction.

[0131] 1.6 Statistical Methods: This study efficiently integrated multiple professional tools such as Origin 2024 (OriginLab), Prism 8 (GraphPad Software), and Excel 2019 (Microsoft) to optimize the entire process of data processing, statistical analysis, and graphical presentation.

[0132] 2. Results

[0133] 2.1 General characteristics of the element fingerprint profile of Peucedanum chinense: This study used a semi-quantitative method to monitor 66 major and trace elements in Peucedanum chinense. All trace elements basically cover metal elements and non-metal elements. PCA, PLSDA and cluster heat map analysis were performed on all elements using the R language package, which included factors from Anhui, Zhejiang, Chongqing, Guizhou and other production areas. The results are as follows Fig.10 As shown, PCA analysis shows that the distances between different production areas are relatively close and not completely separated ( Fig.10 A). Further PLSDA analysis results show that the distances between different production areas are also relatively close, among which the distances between Zhejiang production area and Guizhou, Chongqing and other production areas are relatively far ( Fig.10 B). Cluster heat map analysis results showed that no significant feature clustering effect was observed ( Fig.10C). The preliminary results of the element fingerprint profile study show that it is difficult to find the overall factor characteristics using traditional chemometric methods, and thus it is difficult to reveal the characteristics of the production area. It is necessary to explore more appropriate data identification methods.

[0134] 2.3 Small molecule metabolite profile analysis: This study used high-resolution mass spectrometry to collect the metabolite m / z and response values ​​of all samples of multiple batches of Radix Peucedanum from Anhui, Zhejiang, Chongqing, Guizhou and other production areas. After preliminary standardization and normalization, PCA, OPLSDA and cluster analysis were used to further observe the change profile and characteristics of secondary metabolites ( Fig.11 ). The PCA analysis results show that ( Fig.11 A), the main production areas are relatively chaotic, without showing obvious production area characteristics, and the distance is relatively close. The OPLSDA analysis results show that ( Fig.11 B), Zhejiang production area is well separated from Chongqing, Guizhou and other production areas, and the production area characteristics are clearly observed, while the distances between other production areas are relatively close. The cluster analysis results show that different production areas cannot be well clustered into one category ( Fig.11 C). In summary, this study has preliminarily understood the variation profile of secondary metabolites in Peucedanum samples. Through the production area factor, fewer factor characteristics were observed, and better methods need to be explored to identify factor specificity.

[0135] 3.4 Machine learning for the classification of origin of Peucedanum chinense by integrating metallomics and plant metabolomics: Based on the preliminary chemometric analysis results of non-targeted omics, it is essential to explore more appropriate feature classification methods. Based on the preliminary analysis results of chemometrics, in order to more accurately discover and reveal the quality connotation information of Peucedanum chinense under multiple factors, this study uses machine learning in artificial intelligence as a means, takes Peucedanum chinense element fingerprints, collected metabolite characteristics and the combination of the two as the basic database, and develops and optimizes algorithms to achieve classification modeling under different factors, so as to more accurately and clearly reveal the key characteristics of Peucedanum chinense. In previous studies, this study found that due to the limited data volume of conventional experimental data of traditional Chinese medicine, it is difficult to obtain a suitable model through a suitable algorithm. When dealing with unbalanced data, the training of classification algorithms often leads to poor prediction performance. The model will significantly favor the majority class and ignore the minority class samples that are critical to many practical applications. This makes the model impractical when facing practical problems containing rare but high-priority events. To solve this problem, oversampling technology provides a way to rebalance the categories before model training. By duplicating minority class data points, oversampling achieves a balance in the training data, thereby preventing the algorithm from ignoring those important but scarce categories. Although this method has the risk of overfitting, it can effectively offset the adverse effects of imbalanced learning and enable machine learning models to handle key use cases. In this study, this study innovatively adopted data oversampling methods and ideas to effectively solve the problem of sample size imbalance. The measured values ​​are standardized and normalized, and as a data set for machine learning, they are randomly divided into training sets and test sets, with a training set and a test set ratio of 7:3. Five machine learning algorithms, RF, KNN, Rpa, SVM, and XGB, are used to establish and optimize the model, and AUC and ACC are used to evaluate the performance of the model. The samples cover the four major production areas of Anhui, Zhejiang, Guizhou, and Chongqing.

[0136] 2.4.1 Machine learning classification modeling based on metallomics: This study follows the established machine learning modeling framework and deeply explores the application potential of the response values ​​of 66 elements obtained by semi-quantitative analysis after internal standard correction in the field of non-targeted metallomics. By constructing a preliminary model and evaluating its performance, this study obtained detailed results as shown in Table 1. In the model evaluation phase, this study paid special attention to the AUC value of the test set, and also used the ACC index for comprehensive consideration. The data in Table 1 reveal the significant differences between the models of various production areas. Specifically, the AUC values ​​of the models constructed by different algorithms in the Anhui production area on the test set fluctuated greatly, ranging from 0.78 to 0.93, showing a certain diversity of model performance. The test set AUC range of the Guizhou production area model is between 0.56 and 0.77, indicating that its model prediction ability is relatively robust but still has room for improvement. The performance of the Zhejiang production area is particularly outstanding, with a test set AUC range of up to 0.92 to 0.97, which means that the model of this production area has extremely high prediction accuracy and reliability. The AUC test set value of the Chongqing production area model ranges from 0.85 to 0.94, showing a certain degree of predictive ability but still needs to be optimized. Further analysis of the ROC curve results of all algorithm modeling, such as Fig.12 As shown. The closer the ROC curve is to the upper left corner, the better the performance of the classifier, which means that a high true positive rate is obtained while maintaining a low false positive rate. The characteristics of the ROC curve show that it can help select the best classification threshold. At the same time, combined with the ACC results of the test set for analysis, this study found that the models of the Anhui, Chongqing and Zhejiang production areas performed well in terms of accuracy, significantly better than those of the Guizhou production area. This finding not only verifies the uniqueness of the metal element distribution characteristics of the Anhui, Chongqing and Zhejiang production areas, but also highlights the advantages of the relevant production areas in model construction. In the overall model construction of metallomics, the selection of the optimal algorithm for each production area shows certain regional differences. In the Anhui production area, the RF algorithm has become the first choice for building the optimal model due to its strong learning ability and generalization performance. The optimal algorithm in the Zhejiang production area is KNN. In the Guizhou production area, the SVM algorithm stands out with its unique advantages and becomes the optimal choice for this production area. In the Chongqing production area, the XGB algorithm successfully built the optimal model due to its simple and easy to understand and easy to implement characteristics. In summary, this study has explored the construction and performance evaluation of metallomics models in different production areas through a combination of semi-quantitative analysis and machine learning modeling. The significant differences between the models of different production areas not only reveal the impact of regional element content characteristics on model construction, but also provide a strong basis for the subsequent development of customized models for different production areas.

[0137] Table 1 Evaluation results of classification models of Peucedanum chinense producing areas based on element fingerprints

[0138]

[0139] 2.4.2 Machine learning classification modeling based on plant metabolomics: This study used high-resolution mass spectrometry to obtain mass spectrometry data of secondary metabolites of Peucedanum chinense from different production areas. Based on the distribution profile characteristics of secondary metabolites, five machine learning algorithms were further used to establish a classification model of Peucedanum chinense from different production areas and authenticity characteristics. Considering the complexity of secondary metabolites and the sensitivity of mass spectrometry detection, this study first cleaned the data of the detected molecular ion peaks. 5 The above compound information is the basic data set, and the molecular ion peaks close to the mass spectrum baseline are removed accurately. In addition, in order to further optimize the basic data set and improve its quality, this study specifically performed correlation removal processing on complex metabolite information. The core purpose of this step is to reduce redundant and highly correlated features through careful screening to improve the performance, interpretability, and computational efficiency of the machine learning model, reduce the risk of overfitting, and reduce noise and storage requirements. The analysis results before and after correlation removal are shown in Figure 2. Fig.13 By comparing the correlation characteristics before and after removal, this study found that the correlation between secondary metabolites after removal was lower, which significantly improved the quality of the data.

[0140] This study followed the established machine learning modeling framework, using cleaned secondary metabolite mass spectrometry data as input, and constructed a variety of algorithm models for origin and authenticity classification through meticulous data set division, algorithm selection, and model parameter tuning. The preliminary model construction results based on non-targeted plant metabolomics and their evaluation details are shown in Table 2. When evaluating the model, this study mainly relied on the AUC value of the test set, supplemented by the characteristic analysis of the ROC curve and the accuracy (ACC) for comprehensive consideration. The data in Table 2 show that the test set AUC values ​​of the algorithm models in the Anhui production area ranged from 0.86 to 0.96, showing good classification performance; the AUC value range of the Guizhou production area model was between 0.68 and 0.91; the AUC value range of the Zhejiang production area model was 0.80 to 0.97; and the AUC value of the Chongqing production area model fluctuated between 0.67 and 0.96. Further in-depth analysis of the ROC curves of the models constructed by all algorithms (such as Fig.14 As shown in the figure, combined with the test set accuracy (ACC) for horizontal comparison, this study found that the test set accuracy of Anhui, Chongqing and Zhejiang production areas is relatively high. In the construction of the comprehensive model of metallomics, this study found that the optimal models obtained by different algorithms in each production area are: Anhui production area preferred KNN algorithm model, Zhejiang production area performed best with XGB algorithm model, Guizhou production area SVM algorithm model stood out, and Chongqing production area also took SVM algorithm model as the best.

[0141] Table 2 Evaluation results of classification models of Peucedanum chinense producing areas based on secondary metabolites

[0142]

[0143]

[0144] 2.4.3 Establishment of machine learning aggregation model based on metalomics fusion metabolomics: Based on the classification model results of metalomics and metabolomics, this study found that although the accuracy of the model was greatly improved after data oversampling, the test set AUC and ACC of the metalomics model for some production areas such as Guizhou were low, and the stability between the algorithms based on metabolomics was poor. In order to further improve the accuracy and stability of prediction, reduce the risk of misjudgment, and increase the performance of the model in different scenarios, this study attempted to adopt the idea of ​​aggregation model and use the differences between different models to improve the accuracy and stability of the overall prediction system. Based on the models with higher accuracy obtained by the five algorithms, metalomics and plant metabolomics were aggregated in different production areas. This further improved the prediction effect of the overall model and obtained more reliable prediction results. More importantly, this strategy combined the element fingerprint substances in Peucedanum chinense with the secondary metabolite profile substances for the first time, and gained insight into the origin and authenticity of Peucedanum chinense from the overall material level. The results of the establishment, optimization and evaluation of the aggregation model are shown in Table 3. Comparing the classification model results of metallomics and metabolomics alone (Table 2 and Table 3), this study found that the aggregated model significantly improved the AUC and ACC of each production area. Especially in the Guizhou production area, the test set AUC of the optimal classification model of metallomics before aggregation was 0.77 (SVM), the test set AUC of the optimal classification model of metabolomics was 0.91 (SVM), and the test set AUC established by the aggregated model was 0.94. It can be seen that the aggregated model improves accuracy and improves the robustness of prediction, thereby improving the generalization ability of the model. Compared with other production areas, the same trend is shown. Combined with the ROC curve characteristics of different production areas ( Fig.15 ), the ROCs between the production areas of Anhui, Guizhou, Zhejiang, and Chongqing are all close to the left, and the overfitting phenomenon between the models is small, which further verifies the effectiveness and stability of the aggregation model. The results of the aggregation model in this study show that the strategy of combining element fingerprints and secondary metabolite profiles fully utilizes the differences between different models. By integrating the information of metallomics and metabolomics, it can achieve effective aggregation of information in different production areas, thereby comprehensively improving the accuracy and stability of the overall prediction system and achieving accurate traceability of the production area of ​​​​Pueraria lobata.

[0145] Table 3 Evaluation results of the aggregated model based on metallomics and metabolomics

[0146]

[0147] 3.1 A new strategy for identifying the origin and authenticity of Chinese medicine by integrating metallomics and metabolomics: The quality of Chinese medicinal materials is closely related to their origin. Chinese medicinal materials from different origins differ in terms of growth environment, soil conditions, climate conditions, etc. These differences directly affect the chemical composition and metal element content of Chinese medicinal materials. Metallomics is a new research direction that has emerged in recent years. For Chinese medicine, metallomics is used to study the types, contents and distribution patterns of metal elements in Chinese medicinal materials. Metal elements play an important role in Chinese medicinal materials. They are not only the basic elements that constitute the chemical composition of Chinese medicinal materials, but also participate in the pharmacological effects of Chinese medicinal materials. Through metallomics analysis, the characteristic spectra of metal elements in Chinese medicinal materials can be revealed, providing important information for origin identification. Plant Chinese medicine metabolomics is the study of the types, contents and changes of small molecule metabolites in Chinese medicinal materials. Metabolites are chemical substances produced by Chinese medicinal materials during growth, development and metabolism. They can reflect the physiological state and metabolic pathways of Chinese medicinal materials. Through metabolomics analysis, the metabolic fingerprint of Chinese medicinal materials can be obtained, further revealing the origin characteristics of Chinese medicinal materials. The new strategy of using machine learning metallomics and metabolomics integrated model to identify the origin of traditional Chinese medicine is an innovative and scientific method. It combines the advantages of metallomics, metabolomics and artificial intelligence, including element fingerprints and secondary metabolite profiles, and reveals the characteristics of the origin of traditional Chinese medicine in a more comprehensive, in-depth and intelligent way. Therefore, through the integrated analysis of metallomics and metabolomics, the origin of traditional Chinese medicine can be more accurately identified, thereby providing a scientific basis for the quality control and efficacy evaluation of traditional Chinese medicines. This innovative strategy has broad application prospects in the quality control, efficacy evaluation and protection of traditional Chinese medicine resources of traditional Chinese medicines. This study perfectly applied this strategy to Radix Peucedanum, which can integrate 66 trace elements with rich secondary metabolite information, and realize the traceability of the origin and authenticity of Radix Peucedanum with the help of the intelligent algorithm of the machine learning aggregation model. Therefore, this strategy is comprehensive, accurate and scientific in the traceability of the origin and authenticity of Radix Peucedanum, reflecting the quality connotation and scientific laws of Radix Peucedanum.

[0148] 3.2 Element fingerprints The importance of variables of elements and chemical components in the identification of the origin of Peucedanum chinense: In the field of machine learning and deep learning, the interpretability of models occupies a pivotal position. Although complex models such as deep neural networks and integrated models (such as XGBoost, LightGBM) perform well in predictive accuracy, they are often like "black boxes" and their internal decision-making logic is elusive. In order to solve this problem, SHAP (Shapley Additive ex Planations) came into being. As an effective tool, it reveals and explains the output results of the model by assigning importance values ​​to each feature. SHAP is used to calculate the contribution of each feature to the model prediction results. This study further adopted Shaply additive interpretation to calculate all possible arrangements between variables, and then calculated the average contribution of each variable. In this study, the fastshap package was used to calculate variable importance, and the shapviz package was used to visualize variable importance. The calculated SHAP values ​​of the Anhui production area are shown in the figure. Fig.16 This study found that in the element model of Anhui production area, Mn has the greatest impact on the model prediction results ( Fig.16 A), the distribution trend of detailed element variable SHAP values ​​is shown in the summary Fig.16 As shown in B, the summary graph is used to visualize the importance of features and the direction of influence of feature values. Fig.16 B combines the feature importance and feature effect diagrams to show the distribution of the SHAP value of each feature, especially the quantitative element features such as constant elements Mn, Cr, and Na, and further shows the positive and negative contributions of each feature to the model prediction in each sample. Among the chemical components of the Anhui production area, this study found that peak 16696 had the greatest impact on the model's prediction results ( Fig.16 C), the summary graph further illustrates the positive and negative impacts of peak 16696 in each sample ( Fig.16 D), most samples make positive contributions to the model. Fig.16 In this study, it was found that most of the peak information of 15 chemical components can have a positive impact on the model. Fig.16 In B, it was also found that some of the 15 elements in most samples had a positive impact on the model prediction results. Therefore, based on the interpretation results of the model in this study, the characteristic elements and characteristic chemical components that have an impact can be further combined in the future to explore the possibility of quality control of Peucedanum chinense. At the same time, the analysis results correspond to the superior results of the aggregation model, further explaining and affirming the superiority of the application of the aggregation model in Peucedanum chinense samples.

[0149] In the explanation of the Guizhou production area model ( Fig.17), S and Mg are the two major elements with the greatest impact, and the characteristic importance of peak 13779 is much greater than that of other peaks. Fig.18 ), B, Mg, and peak 19774 have a greater impact on the model prediction results. Fig.19 ), Mg and peak 17630 became the variable features with the greatest impact on the model. In the horizontal comparison of Guizhou, Anhui, Zhejiang, and Chongqing production areas, the SHAP values ​​in the summary diagrams of elements and chemical compositions have their own characteristics, which also well reflects the characteristic variability between samples from production areas.

[0150] The efficacy of Chinese medicine is inseparable from its origin. Accurate traceability of the origin of the production area is helpful to improve quality control and effectiveness. As a commonly used Chinese medicine, the relationship between the quality of Peucedanum chinense and its geographical origin is unknown. Based on this, we innovatively proposed a combined identification model of element fingerprint substances and secondary metabolite profile substances for the identification of the geographical origin of Chinese medicine with Peucedanum chinense as a model. Based on metallomics and metabolomics, the element fingerprint and chemical component profile were integrated. The results showed that compared with chemometrics, machine learning modeling with the help of data oversampling technology can achieve high-precision identification of the production areas of Anhui, Zhejiang, Guizhou and Chongqing. Here, we innovatively adopted an aggregation model to further integrate the optimal models of elements and chemical components to improve the classification accuracy. In addition, the importance of variables gave a professional explanation of the models of elements and chemical components. In summary, for the first time, we have fully understood the scientific connotation of the production area of ​​Peucedanum chinense based on the perspective of integrating element fingerprints and secondary metabolite profile factors based on machine learning, which provides a scientific basis for scientific and reasonable quality supervision and clinical application.

[0151] At the same time, recognizing the indispensability of element fingerprint substances in quality control, we advocate the construction of an innovative comprehensive quality control strategy, that is, to integrate element fingerprints with secondary metabolite profiles to form a more comprehensive and in-depth evaluation system for the quality of medicinal plants. Therefore, a multidimensional quality control strategy integrating element fingerprints and secondary metabolite profiles came into being. The implementation of this comprehensive strategy can improve the scientificity and accuracy of the quality evaluation of traditional Chinese medicine. At present, with the development of artificial intelligence, machine learning, with its powerful computing power, can not only realize the rapid capture of complex system features, but also establish high-precision classification models to achieve the traceability of the origin and authenticity of traditional Chinese medicine. Therefore, with the help of tools and methods in the intelligent era, exploring and optimizing this comprehensive quality control strategy has far-reaching strategic significance for promoting the innovation and development of the quality evaluation system of traditional Chinese medicine.

[0152] This study uses Peucedanum chinense as a model Chinese medicine. Aiming at the problem of origin traceability and quality identification, it innovatively integrates non-targeted and targeted metallomics, plant metabolomics and machine learning technologies to build a set of efficient and accurate analysis strategies. By comprehensively analyzing the elemental fingerprint and secondary metabolite profile of Peucedanum chinense, combined with the construction of intelligent models, it is possible to quickly and accurately identify the origin of Peucedanum chinense. This study can provide new ideas, methods and tools for the quality evaluation of Peucedanum chinense, and set a benchmark for the improvement of the overall quality evaluation system of traditional Chinese medicine and scientific supervision.

[0153] It should be noted that the flowcharts and block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each box in the flowchart or block diagram can represent a module, a program segment, or a part of a code, and the module, program segment, or a part of the code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flowchart, and the combination of boxes in the block diagram and / or flowchart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0154] In general, various example embodiments of the present disclosure may be implemented in hardware or dedicated circuits, software, firmware, logic, or any combination thereof. Certain aspects may be implemented in hardware, while other aspects may be implemented in firmware or software that may be executed by a controller, microprocessor, or other computing device. When various aspects of the embodiments of the present disclosure are illustrated or described as block diagrams, flow charts, or using certain other graphical representations, it will be understood that the boxes, devices, systems, techniques, or methods described herein may be implemented as non-limiting examples in hardware, software, firmware, dedicated circuits or logic, general hardware or controllers or other computing devices, or some combination thereof. Those skilled in the art will clearly understand that, for the convenience and simplicity of description, the specific working processes of the systems, devices, and units described above may refer to the corresponding processes in the aforementioned method embodiments, and will not be repeated here.

[0155] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.

[0156] The units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. In addition, the functional units in the various embodiments of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated units may be implemented in the form of hardware or in the form of software functional units.

[0157] The exemplary embodiments of the present disclosure described in detail above are merely illustrative and not restrictive. It should be understood by those skilled in the art that various modifications and combinations may be made to these embodiments or their features without departing from the principles and spirit of the present disclosure, and such modifications should fall within the scope of the present disclosure.

Claims

1. A method for constructing a Chinese medicine origin tracing model integrating element fingerprints and secondary metabolite profiles, characterized in that: The method comprises: 101, obtaining the CPS response value of each element in the training set samples from different production areas and the first production area label corresponding to the sample; 102, inputting the CPS response value into a first machine learning model to obtain a prediction classification result, comparing it with the first production area label, and optimizing the model according to the comparison result to obtain a metallomics model; 103, obtaining mass spectrometry data of secondary metabolites of training set samples from different production areas and second production area labels corresponding to the samples; the second production area label is consistent with the first production area label; 104, inputting the mass spectrometry data into a second machine learning model to obtain a prediction classification result, comparing it with the second production area label, optimizing the model according to the comparison result, and obtaining a metabolomics model; 105. The optimal model of the metallomics model and metabolomics model in the same production area was selected, and the two optimal models were aggregated to obtain the aggregated model of the same production area as the traceability model of the production area.

2. The method for constructing a Chinese medicine origin tracing model integrating element fingerprints and secondary metabolite profiles according to claim 1, characterized in that: The first production area label or the second production area label includes at least two, which are production area one and production area two; the 105 includes: selecting the best model in the metallomics model of production area one, selecting the best model in the metabolomics model of production area one, and aggregating the two best models to obtain the traceability model of production area one; selecting the best model in the metallomics model of production area two, selecting the best model in the metabolomics model of production area two, and aggregating the two best models to obtain the traceability model of production area two; Optionally, the aggregation method of the aggregation model includes: bootstrap aggregation, boosting, stacking, and mixing.

3. The method for constructing a Chinese medicine origin tracing model integrating element fingerprints and secondary metabolite profiles according to claim 1, characterized in that: When the different production areas include three, the production area label also includes production area three; selecting the best model among the metallomics model of production area three, selecting the best model among the metabolomics model of production area three, and aggregating the two best models to obtain the traceability model of production area three; Optionally, when the different production areas include four, the production area label also includes production area four; the optimal model among the metallomics model of production area four is selected, the optimal model among the metabolomics model of production area four is selected, and the two optimal models are aggregated to obtain the traceability model of production area four; Optionally, the training set samples include: Peucedanum chinense, Peucedanum purpurogenum, Saposhnikovia divaricata; when the training set samples are Peucedanum chinense, the first production area label or the second production area label includes any two or more of the following: Anhui, Zhejiang, Chongqing, Guizhou; Optionally, the CPS response value is the response value after internal standard correction and blank background deduction; Optionally, the elements include constant elements and trace elements; trace elements include metal elements and non-metal elements; Optionally, between 101 and 102, the CPS response value is further preprocessed, and the preprocessing includes: 1011, preliminary standardization processing of dividing the CPS response value by the sampling amount; Optionally, after 1011, the preprocessing further includes: 1012, normalizing the CPS response values ​​of the major elements and trace elements in the training set samples from different production areas; performing dimensionality reduction processing on the normalized CPS response values; the dimensionality reduction processing method includes PCA, OPLS-DA and cluster heat map analysis; Optionally, between 103 and 104, the step further includes performing a cleaning process on the mass spectrum data to reduce redundancy and highly correlated features, to obtain cleaned mass spectrum data; Optionally, the cleaning process includes: 1031, taking the compound information with response values ​​greater than a first threshold as the basic data set, removing the molecular ion peaks that are close to the mass spectrum baseline and accurate; 1032, performing correlation removal processing on the molecular ion peaks that have been removed that are close to the mass spectrum baseline and accurate, to obtain secondary metabolite mass spectrum data after removal.

4. The method for constructing a Chinese medicine origin tracing model integrating element fingerprints and secondary metabolite profiles according to claim 1, characterized in that: Between 101 and 102, there is also an oversampling process for the training set samples, and the oversampling process includes: 100, copying or amplifying minority class data points to balance the number of samples of each category in the training data; the minority class data points include any one or more of the following: specific origin, special ingredient content; optionally, the methods used by the first machine learning model and / or the second machine learning model include: RF, KNN, Rpa, SVM, XGB.

5. A method for tracing the origin of traditional Chinese medicine by integrating element fingerprints and secondary metabolite profiles, characterized in that: The method comprises: 201, obtaining metallomics data and metabolomics data of the sample to be tested; 202, inputting the metallomics data and / or metabolomics data into the traceability model of different production areas for processing, and obtaining a classification result of whether the data belongs to the production area; Optionally, the metallomics data includes CPS response values ​​of any one or more of the following characteristic elements: Mn element, S element, Mg element, B element; Optionally, the metabolomics data includes mass spectrometry data of any one or more of the following chemical component peak information: peak 16696, peak 13779, peak 19774, peak 17630.

6. A method for tracing the origin of traditional Chinese medicine by integrating element fingerprints and secondary metabolite profiles, characterized in that: The method comprises:

301. Obtain the CPS response value of the target characteristic element of the sample to be tested and the mass spectrum data of the target peak information, wherein the target characteristic element includes any one or more of the following: Mn element, S element, Mg element, and B element; the target peak information includes any one or more of the following: peak 16696, peak 13779, peak 19774, and peak 17630; 302. Input the CPS response value of the target characteristic element and the mass spectrum data of the target peak information into the traceability model of different production areas for processing to obtain the classification result of whether it belongs to the production area.

7. The method for tracing the origin of traditional Chinese medicine by integrating element fingerprint and secondary metabolite profile according to claim 5 or 6, characterized in that: The number of the traceability models is consistent with the number of production area labels; the traceability models include at least two, namely, the production area one traceability model and the production area two traceability model; Optionally, the input is processed in the traceability model of different production areas to obtain a classification result of whether it belongs to the production area, including: inputting the metallomics data and / or metabolomics data into the traceability model of production area one, and outputting a yes or no result; if a yes result is output, ending the operation, and obtaining a result that the sample to be tested belongs to production area one; if a no result is output, obtaining a result that the sample to be tested does not belong to production area one, and inputting the metallomics data and / or metabolomics data into the traceability model of production area two, and outputting a yes or no result; if a yes result is output, ending the operation, and obtaining a result that the sample to be tested belongs to production area two; if a no result is output, obtaining a result that the sample to be tested does not belong to production area two.

8. A computer device, characterized in that: The device comprises: a memory and a processor; the memory is used to store a computer program; the processor executes the computer program to implement the steps of the method according to any one of claims 1 to 7.

9. A computer-readable storage medium, characterized in that: A computer program is stored thereon, and when the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method described in any one of claims 1 to 7 are implemented.

Images