Single Chinese herbal medicine multi-component map construction method and system

Abstract

The invention relates to the technical field of traditional Chinese medicine information processing, in particular to a single Chinese herbal medicine multi-component map construction method and system. Obtaining target component categories of a single Chinese herbal medicine, wherein the target component categories comprise active components, symbolic components, auxiliary components and toxic and side components; qualitative and quantitative information of various target components is obtained by combining chromatographic separation, mass spectrum identification, spectral analysis and biological detection technologies; establishing a component association network to reflect synergistic interaction or antagonism among different components; the generated multi-component map of the multi-layer structure comprises a base layer, a correlation layer, a function layer and an application layer, and component composition, content distribution and mutual relation are displayed. Constructing a comprehensive feature library and a difference evaluation system to realize structure speculation and function prediction of unknown components and establish a Chinese herbal medicine quality grading standard; and constructing a component-target-pathway-efficacy association prediction model based on the multi-component map to clarify the overall efficacy of the multi-component synergistic effect. A scientific basis is provided for quality control of traditional Chinese medicines, research and development of new medicines and clinical application.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of traditional Chinese medicine information processing, in particular to a single Chinese herbal medicine multi-component profile construction method and system. BACKGROUND

[0002] As an important part of traditional Chinese medicine system, the effectiveness and safety evaluation of Chinese herbal medicine has always been a research hotspot. With the development of analytical techniques, the research on the components of Chinese herbal medicine has gradually shifted from single-component mode to multi-component comprehensive analysis mode. Traditional research on Chinese herbal medicine adopts the reductionist approach of "one herb-one component-one target", and studies are carried out by isolating and purifying single active components. In recent years, the application of modern analytical methods such as high performance liquid chromatography-mass spectrometry (HPLC-MS), gas chromatography-mass spectrometry (GC-MS) and nuclear magnetic resonance (NMR) has made it possible to simultaneously detect multiple components of Chinese herbal medicine. The progress of data processing technology, especially the data processing system under G06F, provides technical support for the integration, analysis and visualization of Chinese herbal medicine component information. Existing research shows that the efficacy of Chinese herbal medicine is often the result of the synergistic action of multiple components, rather than a single component.

[0003] However, the existing research on the components of Chinese herbal medicine has obvious shortcomings. First, most researches still focus on a few known active components, and insufficient attention is paid to auxiliary components, toxic and side components, making it difficult to fully reflect the overall characteristics of Chinese herbal medicine; second, the existing technology does not adequately analyze the interactions between components, especially the lack of systematic research on synergistic or antagonistic effects; third, the detection methods for components are single, which cannot cover the diverse components of Chinese herbal medicine, resulting in incomplete component information; fourth, component data usually exists in a scattered form, lacking effective integration means, and it is impossible to establish a correlation network between components; finally, the visualization of Chinese herbal medicine component information is limited, which makes it difficult to intuitively display the composition, content distribution and mutual relationship of components, and is not conducive to the understanding and application of the overall properties of Chinese herbal medicine. These problems seriously hinder the in-depth development of Chinese herbal medicine research and the scientificity of clinical application. SUMMARY

[0004] In view of the problems existing in the prior art, the present application is proposed.

[0005] Therefore, the problem to be solved by the present application is how to solve the technical problem of multi-component comprehensive analysis and visualization of single Chinese herbal medicine. By establishing a complete component analysis system, the present application realizes the comprehensive detection of multiple types of components of Chinese herbal medicine; by constructing a component correlation network, the present application reveals the interaction between different components; by generating multi-level component profiles, the present application intuitively displays the overall characteristics of Chinese herbal medicine, and provides a scientific basis for the quality control of traditional Chinese medicine, the research and development of new drugs and clinical application.

[0006] To solve the above technical problems, the present application provides the following technical solutions: In the first aspect, the present application provides a single Chinese herbal medicine multi-component mapping construction method, which comprises determining target component categories of a single Chinese herbal medicine, wherein the target component categories include active components, marker components, auxiliary components and toxic and side components; Qualitative and quantitative information of each type of target component in the single Chinese herbal medicine is obtained by using multiple detection technologies; An ingredient correlation network of the single Chinese herbal medicine is established, which reflects synergistic or antagonistic effects between different components; A multi-component map of the single Chinese herbal medicine is generated, which shows component composition, content distribution and mutual relationship.

[0007] As a preferred scheme of the single Chinese herbal medicine multi-component mapping construction method of the present application, wherein: the target component categories further include: Chemical fingerprint characteristic components, function dominant components and structurally similar component groups.

[0008] As a preferred scheme of the single Chinese herbal medicine multi-component mapping construction method of the present application, wherein: the multiple detection technologies include: A combination of chromatographic separation technology, mass spectrometry identification technology, spectral analysis technology and biological detection technology.

[0009] As a preferred scheme of the single Chinese herbal medicine multi-component mapping construction method of the present application, wherein: the method further comprises: Constructing a single Chinese herbal medicine multi-component comprehensive feature library to record physicochemical properties, structural information and biological activity of each component; Based on the comprehensive feature library, structure speculation and function prediction are performed on unknown components.

[0010] As a preferred scheme of the single Chinese herbal medicine multi-component mapping construction method of the present application, wherein: the method further comprises: Analyzing variation rules of single Chinese herbal medicine multi-components under different production places, different harvesting times and different processing technology conditions; Establishing a component difference evaluation system to determine a Chinese herbal medicine quality grading standard.

[0011] As a preferred scheme of the single Chinese herbal medicine multi-component mapping construction method of the present application, wherein: the method further comprises: Based on the multi-component map of the single Chinese herbal medicine, a correlation prediction model of components-targets-pathways- efficacy is established; A Chinese herbal medicine action mechanism visualization network is generated to clarify the overall pharmacodynamic effect of multi-component synergistic action.

[0012] As a preferred scheme of the single Chinese herbal medicine multi-component pattern construction method, the method comprises the following steps: In a second aspect, the embodiments of the present application provide a single Chinese herbal medicine multi-component pattern construction system, which comprises a component category determination module for determining a target component category of a single Chinese herbal medicine, wherein the target component category comprises active components, marker components, auxiliary components and toxic and side components; A component information acquisition module is configured to acquire qualitative and quantitative information of each type of target component in the single Chinese herbal medicine by using multiple detection technologies; An association network establishment module is configured to establish a component association network of the single Chinese herbal medicine, wherein the association network reflects synergistic or antagonistic effects between different components; A multi-component pattern generation module is configured to generate a multi-component pattern of the single Chinese herbal medicine, wherein the pattern shows component composition, content distribution and mutual relationship.

[0013] In a third aspect, the embodiments of the present application provide a computer device, which comprises a memory and a processor, and the memory stores a computer program, wherein the computer program instructions are executed by the processor to realize the steps of the single Chinese herbal medicine multi-component pattern construction method according to the first aspect of the present application.

[0014] In a fourth aspect, the embodiments of the present application provide a computer readable storage medium, which stores a computer program, wherein the computer program instructions are executed by the processor to realize the steps of the single Chinese herbal medicine multi-component pattern construction method according to the first aspect of the present application.

[0015] The single Chinese herbal medicine multi-component pattern construction method and system provided by the present application have significant technical advantages and application value. By establishing a multi-dimensional component category system including active components, marker components, auxiliary components and toxic and side components, the limitations of traditional research focusing on only a few active components are broken through, and the overall characteristics of the components of Chinese herbal medicines are comprehensively grasped. The optimal combination of multiple detection technologies such as chromatographic separation, mass spectrometry identification, spectral analysis and biological detection solves the technical problem that a single technology cannot comprehensively cover the structural diversity of components of Chinese herbal medicines, and improves the comprehensiveness and accuracy of component detection.

[0016] The component association network constructed based on multi-source data reveals the synergistic or antagonistic effect relationships between different components in Chinese herbal medicines, and provides a scientific basis for understanding the overall action characteristics of Chinese herbal medicines in terms of "multi-component, multi-target and multi-pathway". The multi-component pattern with multi-level structure presents the complex component information of Chinese herbal medicines in a systematic and structured manner in the form of a basic layer, an association layer, a function layer and an application layer, so that researchers can form a complete technical chain from composition analysis to efficacy evaluation, greatly improving the understandability and application value of the data.

[0017] The constructed comprehensive feature library can not only systematically record known component information, but also make structural speculation and functional prediction on unknown components, expand the information dimension of the atlas. The component difference evaluation system provides a scientific basis for the establishment and grading of Chinese herbal medicine quality standards, and promotes the standardized development of the industry. The component-target-pathway-effect correlation prediction model based on multi-component atlas clarifies the molecular mechanism of multi-component synergistic action of Chinese herbal medicine, and provides theoretical support for efficacy evaluation and new drug research and development.

[0018] In summary, the present application realizes the technical leap from single component to multi-component system, from static analysis to dynamic correlation, from composition description to efficacy interpretation, provides a new technical platform for traditional Chinese medicine quality control, efficacy mechanism research and new drug development, and promotes the development of traditional Chinese medicine modernization research in the direction of systematization, precision and intelligentization. BRIEF DESCRIPTION OF DRAWINGS

[0019] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can also be obtained by those skilled in the art without creative labor.

[0020] Fig. 1 Flow chart of single Chinese herbal medicine multi-component atlas construction method; Fig. 2 Computer device diagram of single Chinese herbal medicine multi-component atlas construction method; Fig. 3 Single Chinese herbal medicine multi-component atlas construction method. DETAILED DESCRIPTION

[0021] In order to make the above-mentioned purposes, features and advantages of the present application more obvious and easy to understand, the specific embodiments of the present application will be described in detail below with reference to the drawings.

[0022] In the following description, many specific details are set forth in order to provide a thorough understanding of the present application, but the present application can also be implemented in other ways different from the description, and those skilled in the art can make similar generalizations without departing from the connotation of the present application, therefore the present application is not limited by the specific embodiments disclosed below.

[0023] Secondly, the "one embodiment" or "embodiment" referred to herein means that the specific features, structures or characteristics can be included in at least one implementation of the present application. "In one embodiment" appearing in different places in the specification does not mean the same embodiment, nor is the embodiment alone or selectively excluded from other embodiments.

[0024] Embodiment 1 Reference Figs. 1-2 For a first embodiment of the present application, the embodiment provides a single Chinese herbal medicine multi-component mapping construction method, comprising, S100: determining the target component categories of a single Chinese herbal medicine, the target component categories including active ingredients, marker ingredients, auxiliary ingredients, and toxic and side ingredients; S200: acquiring qualitative and quantitative information of each type of target component in the single Chinese herbal medicine by using multiple detection technologies; S300: establishing a component association network of the single Chinese herbal medicine, the association network reflecting synergistic or antagonistic effects between different components; S400: generating a multi-component map of the single Chinese herbal medicine, the map showing the component composition, content distribution, and mutual relationship.

[0025] It should be noted that as an important part of the traditional medicine system, the effectiveness and safety evaluation of Chinese herbal medicine has always been a research hotspot. However, the existing component research of Chinese herbal medicine has many technical limitations: first, most researches still focus on a few known active ingredients, and insufficient attention is paid to auxiliary ingredients, toxic and side ingredients, etc., making it difficult to fully reflect the overall characteristics of Chinese herbal medicine; second, the existing technology does not sufficiently analyze the interaction between components, especially the lack of systematic research on synergistic or antagonistic effects; third, the component detection method is single, which cannot cover the structurally diverse components of Chinese herbal medicine, resulting in incomplete component information; fourth, the component data usually exists in a scattered form, lacking effective integration means, and it is difficult to establish an association network between components; finally, the visualization expression means of Chinese herbal medicine component information is limited, and it is difficult to directly show the component composition, content distribution, and mutual relationship. These problems seriously restrict the in-depth development of Chinese herbal medicine research and the scientificity of clinical application. Through the implementation steps of S100-S400, the present application constructs a complete Chinese herbal medicine multi-component mapping construction system, which can systematically and comprehensively analyze each type of component in Chinese herbal medicine, establish the interaction relationship between components, and directly show the overall characteristics of Chinese herbal medicine in an intuitive and visual manner, thereby providing a scientific basis for traditional Chinese medicine quality control, new drug research and development, and clinical rational drug use.

[0026] It should be noted that as an important part of traditional medicine system, the effectiveness and safety evaluation of Chinese herbal medicine has been a research hotspot. However, the existing Chinese herbal medicine composition research has many technical limitations: first, most of the researches still focus on a few known active ingredients, and insufficient attention is paid to auxiliary ingredients, toxic and side ingredients, etc., which is difficult to fully reflect the overall characteristics of Chinese herbal medicine; second, the existing technology is not deep enough in analyzing the interaction between ingredients, especially the lack of systematic research on synergistic or antagonistic effects; third, the single detection method of ingredients cannot cover the structurally diverse ingredients of Chinese herbal medicine, resulting in incomplete ingredient information; fourth, the ingredient data usually exists in a scattered form, lacking effective integration means, and unable to establish the correlation network between ingredients; finally, the visualization expression means of Chinese herbal medicine composition information is limited, and it is difficult to intuitively display the composition, content distribution and mutual relationship of ingredients. These problems seriously restrict the in-depth development of Chinese herbal medicine research and the scientificity of clinical application. Through the implementation steps of S100-S400, a complete Chinese herbal medicine multi-ingredient spectrum construction system is constructed, which can systematically and comprehensively analyze various ingredients in Chinese herbal medicine, establish the interaction relationship between ingredients, and display the overall characteristics of Chinese herbal medicine in an intuitive and visual way, providing a scientific basis for traditional Chinese medicine quality control, new drug research and development and clinical rational drug use.

[0027] Embodiment 2 Reference Figs. 1-3 , the second embodiment of the present application.

[0028] In the embodiments of the present application, the target ingredient categories of a single Chinese herbal medicine are determined in step S100, and the target ingredient categories include active ingredients, marker ingredients, auxiliary ingredients and toxic and side ingredients, including the following steps A1-A2: A1: The target ingredient categories further include: Chemical fingerprint characteristic ingredients, function dominant ingredients, and structurally similar ingredient families.

[0029] Specifically, in A1, the chemical fingerprint characteristic ingredients refer to compounds that can stably appear and have characteristic retention time and mass spectrum fragment pattern in chromatography-mass spectrometry analysis, which are usually used for identification and authenticity identification of Chinese herbal medicine. For example, the chemical fingerprint characteristic ingredients of ginseng mainly include ginsenoside Rg1 (retention time about 10.5 minutes, m / z 801.4931), ginsenoside Re (retention time about 9.8 minutes, m / z 947.5513) and ginsenoside Rb1 (retention time about 16.2 minutes, m / z 1109.6100) and the like, which form a typical fingerprint pattern in UPLC-QTOF-MS analysis. In practical application, the matching degree of the sample to be tested and the standard fingerprint is evaluated by calculating the similarity (such as cosine similarity), and generally the similarity ≥0.85 is considered as a qualified sample.

[0030] Function-leading components refer to the key component groups that contribute most to a specific pharmacological effect of a Chinese herb. Such components are usually determined through active separation tracking experiments, i.e., through bioactivity-oriented separation and purification process, to track and confirm the key components in the main active fraction. For example, the cardiovascular protective effect of Danshen is mainly contributed by liposoluble and water-soluble components such as tanshinone IIA (regulating calcium channel, IC50=15 μM), cryptotanshinone (inhibiting platelet aggregation, EC50=27.5 μM), and salvianolic acid B (antioxidant effect, DPPH free radical scavenging rate 89.3%), which together constitute the function-leading component group of Danshen. In the construction of multi-component maps, function-leading components need to be highlighted and their content and activity data need to be recorded in detail.

[0031] Structurally similar component groups refer to component groups with the same or similar parent structure, only differing in substituents or stereo configuration. Such components usually have similar biosynthetic pathways and pharmacological activities, and often exist in the form of "groups" in Chinese herbs and exert synergistic effects. For example, the flavonoid component group in Huangqin includes baicalin (7-O-glucoside of 5,6,7-trihydroxyflavone), wogonoside (7-O-glucoside of 5,7,2',6'-tetrahydroxyflavone), and bacaleurin (5,6,7-trihydroxyflavone), which share the same flavone skeleton and exhibit similar anti-inflammatory and antioxidant properties, but differ in activity intensity and pharmacokinetic properties due to different substituents. In multi-component maps, these structurally similar component groups are usually classified and displayed through structural clustering algorithms (such as Tanimoto coefficient to calculate similarity, threshold set to 0.85).

[0032] A2: Determination methods and standards for different categories of components include: Active components are determined based on pharmacological screening experiment results, such as IC50≤50 μM or in vivo effective dose≤20 mg / kg; Marker components are determined based on specificity evaluation and stability testing, with specificity≥0.9 and batch-to-batch variation coefficient≤5%; Auxiliary components are determined based on synergistic effect enhancement experiments, with a significant synergistic effect identified as a synergistic coefficient CI value≤0.8; Toxic and side components are determined based on toxicology evaluation, with cytotoxicity CC50≤100 μM or animal LD50≤500 mg / kg.

[0033] Specifically, in A2, the determination of active ingredients adopts a hierarchical screening strategy. First, a preliminary activity database is established through in vitro screening experiments (such as enzyme inhibition experiments, cell function experiments, etc.), and the screening criteria are generally IC50≤50 μM or EC50≤30 μM; second, in vivo pharmacodynamic verification is performed on the positive components of the preliminary screening, such as determining the effective dose in the corresponding disease model (hypertension model, inflammation model, etc.), and the standard is generally ED50≤20 mg / kg; finally, comprehensive evaluation is carried out in combination with pharmacokinetic parameters (such as oral bioavailability ≥10%, half-life ≥2 hours) to determine the final list of active ingredients. For example, in the screening of antibacterial active ingredients of Coptis chinensis, berberine (MIC = 125 μg / mL) and palmatine (MIC = 250 μg / mL) were first screened out through the inhibition zone experiment; then their in vivo antibacterial activity was verified in a mouse infection model (effective doses were 15 mg / kg and 28 mg / kg, respectively); and finally, berberine was determined as the main active ingredient in combination with the oral bioavailability data (5.2% and 3.8%, respectively).

[0034] The determination of a marker component needs to consider the specificity, stability and quantitative ease of three aspects. The specificity evaluation is calculated by comparing the component composition of the same genus and different species of Chinese herbal medicine, and the specificity index (SI = the number of specific components of this species / the total number of components) is required to be SI≥0.9; the stability test is analyzed by analyzing the content variation of the target component in different batches of samples, and the batch variation coefficient RSD is required to be≤5%; the quantitative ease of investigation includes factors such as chromatographic behavior (peak shape, separation degree), detection sensitivity (signal-to-noise ratio S / N≥10) and linear range (r 2 ≥0.999). For example, the marker component of Astragalus membranaceus, astragaloside A, has a specificity index of 0.95 in different species of Astragalus, and a content RSD of 3.2% in the analysis of 30 batches of samples, and has good chromatographic behavior (Rs≥1.5) and detection characteristics (LOQ=0.1 μg / mL, linear range 0.5-100 μg / mL, r 2 =0.9998) in the UPLC-DAD method, so it is determined as the marker component of Astragalus membranaceus.

[0035] The determination of auxiliary components is mainly based on synergistic effect experiments. The isobologram or combination index method is used to evaluate the synergistic effect of two or more components when used together. When the combination index CI value is ≤0.8, it is considered to have significant synergistic effect, and the related components are determined as auxiliary components. For example, when Danshensu and tanshinone IIA in Danshen are used together, they show significant synergistic effect in anti-platelet aggregation (CI=0.62), so Danshensu is determined as an auxiliary component of tanshinone IIA in the application scenario of Danshen with anti-thrombus as the main efficacy. In addition, auxiliary components also include components that can improve the pharmacokinetic characteristics of the main active components, such as increasing absorption (increasing Cmax≥50%), prolonging half-life (extending t1 / 2≥2 times) or reducing metabolism (decreasing clearance rate CL≥30%) to enhance the bioavailability of active components.

[0036] The determination of toxic and side components requires comprehensive toxicological evaluation results. First, potential toxic components are preliminarily screened by in vitro cytotoxicity tests (such as MTT method, LDH release method, etc.), with a standard of CC50≤100 μM; second, the safety dose range of the components is evaluated by acute toxicity test, with a standard of animal LD50≤500 mg / kg; finally, the long-term drug safety is evaluated by sub-chronic or chronic toxicity test, including the effects on liver and kidney function, blood system and reproductive system, etc. For example, aconitine series components in Fuzi are determined as the main toxic and side components through cytotoxicity test (CC50=32 μM) and acute toxicity test (LD50=1.8 mg / kg), which need to be highlighted and monitored in the multi-component spectrum (safety limit value≤0.2 mg / g).

[0037] It should be noted that the reasonable determination of target component categories is the basis of multi-component spectrum construction, which directly affects the scientificity and application value of the spectrum. Traditional Chinese herbal medicine research mainly focuses on a few main active components, which fails to fully reflect the complex composition characteristics of Chinese herbal medicine. Through the multi-dimensional component classification system proposed in the present application, the activity contribution, quality characteristics, interaction and safety of Chinese herbal medicine can be fully considered, providing clear direction guidance for the detection and correlation network construction of subsequent target components. In addition, the establishment of determination methods and standards for different categories of components makes the component classification process more scientific and objective, avoiding the randomness and one-sidedness of component selection in traditional research.

[0038] In an alternative embodiment, a knowledge graph-based intelligent active ingredient classification recommendation system can be employed. This system integrates traditional Chinese medicine ancient literature, modern chemical component database and pharmacological research literature, and constructs a knowledge graph containing multi-level relationships of "Chinese herbal medicine-component-property-efficacy". The system learns the complex association patterns in the knowledge graph through graph neural network algorithm, and can automatically recommend the potential target ingredient classification and its representative components of a certain Chinese herbal medicine when the name of the Chinese herbal medicine is input. For example, when "huangqin" is input, the system will recommend flavonoids represented by baicalin and wogonoside, structurally similar component groups represented by 5,7,2'-trihydroxyflavone, and potential synergistic relationship with iridoid glycoside components. This system greatly improves the efficiency and comprehensiveness of target ingredient classification determination, and is particularly suitable for Chinese herbal medicines with less chemical component research.

[0039] In another alternative embodiment, a multi-source omics data-driven target ingredient discovery platform can be employed. This platform integrates metabolomics, transcriptomics and proteomics technologies, and compares the multi-omics differences between effective and ineffective parts of Chinese herbal medicines, and between active and inactive extracts, to mine potential key ingredient markers from massive data. The platform uses multivariate statistical methods and machine learning algorithms such as orthogonal partial least squares discriminant analysis (OPLS-DA) and random forest (Random Forest) to screen ingredient markers with significant differences (VIP>1.5, p<0.01) and activity correlation (|r|>0.8). For example, by comparing the metabolomics data of different polarity extracts and the corresponding antioxidant activity data, 5 previously unattended phenolic acid compounds in danshen were found to have strong correlation with antioxidant activity. This platform can discover new active ingredients that are difficult to identify by traditional methods, and enrich the diversity of the target ingredient library.

[0040] In the embodiments of the present application, a variety of detection techniques are used to obtain qualitative and quantitative information of various target ingredients in a single Chinese herbal medicine in step S200, including the following steps B1-B2: B1: The variety of detection techniques includes: A combination of chromatographic separation technology, mass spectrometry identification technology, spectroscopic analysis technology and biological detection technology.

[0041] Specifically, in B1, chromatographic separation techniques are the basic means to achieve the separation of complex Chinese herbal medicine ingredients, mainly including the following: (1) high performance liquid chromatography (HPLC), equipped with C18, C8, phenyl, amino and other different selectivity chromatographic columns, the mobile phase usually uses methanol / acetonitrile-water system, gradient elution, flow rate 0.8-1.2 mL / min, suitable for the separation of most polar and medium polar components; (2) ultra performance liquid chromatography (UPLC), using high efficiency chromatographic column with particle size <2 μm, system pressure up to 15000 psi, flow rate usually 0.3-0.5 mL / min, 5-10 times higher separation efficiency than traditional HPLC, especially suitable for rapid separation of low content, similar structure components; (3) gas chromatography (GC), equipped with HP-5MS, DB-WAX and other capillary columns, temperature program setting range usually 50-300℃, heating rate 3-20℃ / min, suitable for volatile and heat stable components analysis; (4) hydrophilic interaction liquid chromatography (HILIC), using silica gel, amino and other hydrophilic stationary phase, mobile phase is acetonitrile-water-buffer system, suitable for polar compounds and hydrophilic components such as sugars, amino acids and other separation; (5) capillary electrophoresis (CE), including zone capillary electrophoresis (CZE) and micellar electrokinetic capillary chromatography (MEKC) two main modes, buffer pH range 2.0-9.0, voltage usually 15-30 kV, suitable for charged compounds and ion type components high efficiency separation.

[0042] Mass spectrometry identification techniques are the main means to realize the accurate structure identification of ingredients, mainly including: (1) quadrupole-time of flight mass spectrometry (Q-TOF-MS), mass accuracy <5ppm, resolution >30000, through accurate molecular weight and characteristic fragment spectrum to realize the qualitative analysis of ingredients, suitable for the structure identification of medium molecular weight (100-2000 Da) compounds; (2) triple quadrupole mass spectrometry (QQQ-MS), through multiple reaction monitoring mode (MRM) to realize high sensitivity quantitative analysis, detection limit usually reaches ng / mL level, wide linear range (3-4 orders of magnitude), suitable for high sensitivity quantitative analysis of known structure ingredients; (3) ion trap mass spectrometry (IT-MS) or orbitrap mass spectrometry (Orbitrap-MS), can carry out MSn (n≥3) cascade spectrum acquisition, which is helpful for the in-depth analysis of complex structure, the resolution can reach 240000, suitable for the accurate structure confirmation of complex structure ingredients; (4) matrix assisted laser desorption ionization-time of flight mass spectrometry (MALDI-TOF-MS), suitable for the molecular weight determination of macromolecules such as polysaccharides, proteins and other macromolecules, the mass range can reach more than 300 kDa. Ion source, mainly using electrospray ionization (ESI), atmospheric pressure chemical ionization (APCI) and atmospheric pressure photoionization (APPI) and other soft ionization techniques, respectively suitable for ionization of polar compounds, medium polar compounds and low polar compounds.

[0043] Spectroscopic analysis techniques are important tools for assisting in structural confirmation and quantitative analysis, mainly including: (1) Ultraviolet-Vis spectroscopy (UV-Vis), with a wavelength range of 190-800 nm, mainly used for the qualitative and quantitative analysis of compounds containing conjugated systems, such as flavonoids and anthraquinones; (2) Infrared spectroscopy (IR), with a wavenumber range of 4000-400 cm-1, mainly used to detect functional groups in compounds, such as hydroxyl, carbonyl, and ester groups; (3) Nuclear magnetic resonance spectroscopy (NMR), including one-dimensional spectroscopy (1H-NMR). (3) 13C-NMR and two-dimensional spectra (HSQC, HMBC, COSY, NOESY, etc.) provide the most comprehensive structural information and are the gold standard for confirming the structure of complex natural products; (4) Circular dichroism (CD), with a wavelength range of 180-700nm, is used to determine the stereoconfiguration and absolute configuration of compounds, and is particularly suitable for compounds with many chiral centers; (5) Fluorescence spectroscopy (FL), through characteristic excitation and emission wavelengths, is used for high-sensitivity detection of fluorescent compounds, with detection limits up to the pg level.

[0044] Biological detection techniques are a direct means of evaluating the activity of components, mainly including: (1) enzyme activity analysis, such as measuring the inhibitory activity of compounds on key enzymes such as COX-2, α-glucosidase, and acetylcholinesterase, usually using colorimetric or fluorescence methods to measure changes in substrate conversion rate; (2) cell function assays, including cell proliferation / inhibition assays (MTT method), cell apoptosis assays (flow cytometry), ROS production assays (DCFH-DA probe method), and inflammatory factor expression assays (ELISA or RT-PCR method); (3) receptor binding assays, using radioligand competition or surface plasmon resonance (SPR) techniques to measure the binding ability of compounds to specific receptors; (4) reporter gene systems, using fluorescent or luminescent reporter genes controlled by specific transcription factor response elements to detect the regulatory effect of compounds on specific signaling pathways; (5) in vitro antioxidant activity assays, including DPPH free radical scavenging assay, ABTS free radical scavenging assay, and FRAP reducing power assay, to evaluate the antioxidant activity of components.

[0045] B2: Strategies for optimizing and combining multiple detection technologies include: Construct two-dimensional or multidimensional chromatographic separation systems to improve the total peak capacity of complex samples; Using online or offline colorimetric-mass spectrometry technology, one-step separation and identification can be achieved; Develop bioactivity-guided separation and analysis procedures to improve the efficiency of active ingredient discovery; Establish a multi-index quantitative method for component content to ensure the accuracy and comprehensiveness of the quantitative results.

[0046] Specifically, in B2, the bidirectional or multidimensional chromatographic separation system is an effective strategy to solve the complexity of Chinese herbal medicine ingredients. A typical two-dimensional liquid chromatography system includes a combination of orthogonal separation mechanisms, such as hydrophobicity (RP) x ion exchange (IEX), hydrophobicity (RP) x hydrophilicity (HILIC), or size exclusion (SEC) x hydrophobicity (RP), etc. The system design needs to consider the orthogonality of two-dimensional separation (separation by different characteristics such as hydrophobicity, charge, molecular weight, etc.), flow phase compatibility (realized by dilution, pH adjustment or online SPE), and modulation time (usually 1 / 3-1 / 5 of the elution peak width of the first dimension). For example, in the analysis of ginsenoside components, a HILIC x RP-LC two-dimensional system was used, the first dimension used an XAmide hydrophilic column (4.6 x 250 mm, 5 μm), the mobile phase was acetonitrile-water-ammonium formate buffer, gradient elution 90%-60% acetonitrile, flow rate 0.7 mL / min; the second dimension used a C18 reversed-phase column (2.1 x 100 mm, 1.7 μm), the mobile phase was acetonitrile-water-formic acid, gradient elution 5%-95% acetonitrile, flow rate 0.3 mL / min. The total peak capacity of the system reached more than 2000, successfully separated and identified 58 saponin components in ginseng, far exceeding the separation capacity of conventional one-dimensional chromatography.

[0047] LC-MS, GC-MS, CE-MS, etc. are required to effectively remove the mobile phase (LC uses electrospray interface, GC uses EI or CI interface) and make the separation components compatible with mass spectrometry (flow rate, salt concentration, etc.). The offline coupling mode is mainly aimed at systems that are difficult to directly compatible, such as LC-NMR, which usually adopts peak collection for offline analysis. In practical applications, the following optimization strategies are often used for the analysis of Chinese herbal medicine samples by LC-MS: (1) For components with large polarity differences, gradient elution program is used, usually the initial organic phase ratio is 5%-10%, the final point is 95%-100%, and the gradient time is set to 30-60 min according to the complexity of the components; (2) For isomers, select chromatographic columns with stereoselectivity, such as chiral columns or steric selective columns; (3) For trace components, use large volume injection (30-100 μL) combined with online enrichment technology, or use selective ion scanning mode (SIM or MRM) to improve detection sensitivity; (4) For unstable components, control the column temperature (usually ≤30℃) and add stabilizers (such as antioxidants 0.1% ascorbic acid) to the mobile phase. For example, in the analysis of Angelica sinensis components, a UPLC-Q-TOF-MS system was used, the chromatographic conditions were ACQUITY UPLC HSS T3 column (2.1×100mm, 1.8μm), the mobile phase was 0.1% formic acid aqueous solution (A) and 0.1% formic acid acetonitrile (B), the gradient elution program was 0-1min, 5%B; 1—3min, 5%~20%B; 3—18min, 20%~55%B; 18—23min, 55%~100%B, the flow rate was 0.4mL / min, the column temperature was 35℃, the injection volume was 5μL; the mass spectrometry conditions were ESI source, positive and negative ion mode alternately scanned, scan range m / z 100-1500, capillary voltage 3.0kV, cone hole voltage 40V, collision energy 5-45eV (used to obtain primary and secondary mass spectra). This method successfully identified 87 compounds in Angelica sinensis, including ferulic acid, phthalide, coumarin and polysaccharide, etc.

[0048] Bioactivity-guided fractionation is an effective strategy to improve the efficiency of bioactive component discovery. The procedure usually includes the following steps: (1) establish a micro- or high-throughput bioactivity detection system, such as 96-well plate enzyme inhibition assay, cell viability assay, etc.; (2) perform preliminary separation of the crude extract, such as liquid-liquid extraction, column chromatography, etc., to obtain multiple fractions; (3) evaluate the bioactivity of each fraction to determine the active fractions; (4) perform further separation of the active fractions, such as preparative chromatography, repeated column chromatography, etc.; (5) combine bioactivity determination to track the active sub-fractions until the single active component is isolated. In this procedure, online activity detection technology can be used to improve efficiency, such as HPLC post-column derivatization and enzyme reaction system, which directly introduces the separated fractions into the reaction pool containing the target enzyme and substrate, and evaluates the enzyme inhibition activity of each fraction in real time by detecting the changes in enzyme reaction products. For example, in the study of anti-inflammatory components of Scutellaria baicalensis, an HPLC-COX-2 enzyme inhibition activity online detection system was used to mix the separated fractions with pre-prepared COX-2 enzyme and arachidonic acid substrate, and the anti-inflammatory activity was evaluated by detecting the reduction rate of PGE2 product, successfully discovering four previously unreported anti-inflammatory active components.

[0049] Multi-index quantitative method of component content is the key to ensure the comprehensiveness of quality evaluation of Chinese herbal medicines. This method usually uses the strategy of multi-wavelength or multi-detector combination to simultaneously quantify multiple types of structural components. The specific implementation includes: (1) selecting appropriate extraction methods to ensure the extraction efficiency of various types of target components, usually using ultrasonic-assisted extraction (power 300-500 W, time 30-60 min) or pressurized liquid extraction (temperature 60-100 ℃, pressure 1500 psi, standing time 5-10 min) and other techniques; (2) establishing chromatographic conditions suitable for simultaneous separation of multiple components, usually using long columns (250 mm), small particle size fillers (3-5 μm), and multi-stage gradient elution programs; (3) selecting appropriate detection wavelengths or detectors for components with different structural characteristics, such as 330-360 nm for flavonoids, 410-430 nm for anthraquinones, and FID detector for volatile oils; (4) using characteristic ion multiple reaction monitoring (MRM) mode for high sensitivity and high specificity quantitative analysis, selecting the best precursor ion → product ion pair for each target component, such as Scutellarein m / z 447.1→271.0, Wogonoside m / z 463.1→287.1. In terms of methodological validation, the linear range (usually requiring r 2>0.999), detection limit (S / N > 3), quantification limit (S / N > 10), precision (RSD < 2%), repeatability (RSD < 5%) and recovery (90%-110%) and other parameters. For example, in the quality control study of Coptis chinensis, an UPLC-DAD-QTRAP-MS / MS based simultaneous quantitative method for 11 alkaloid components was established, with a linear range of 0.5-200 μg / mL, a detection limit of 0.05-0.2 μg / mL, an intra-day and inter-day precision RSD of <3.5%, and a recovery of 98.2%-102.6%. The accurate quantification of multiple components in Coptis chinensis was successfully achieved.

[0050] It should be noted that the complexity and diversity of Chinese herbal medicine components determine that a single detection technology cannot meet the demand of comprehensive analysis. Through the optimized combination and use of multiple detection technologies, comprehensive detection of different polarity, different structural types and different content components can be achieved, providing comprehensive data support for the construction of multi-component chromatogram. Chromatographic separation technology solves the problem of component separation, mass spectrometry identification technology solves the problem of component structure confirmation, spectral analysis technology provides complementary structural information, and biological detection technology directly evaluates the activity characteristics of the components. The combination strategy of multiple technologies not only saves analysis time, but also improves data integrity and reliability, providing strong technical support for the multi-component system research of Chinese herbal medicine.

[0051] In an alternative embodiment, a high-throughput sample pretreatment automation platform can be used. The platform integrates automatic equipment for sample grinding, weighing, extraction, filtration and concentration, and realizes full-process automation operation through a central control system. The system includes an automatic grinding module (using a cryogenic grinding technology, with a temperature control below-20°C to prevent degradation of heat-sensitive components), a precise automatic weighing module (with a precision of ±0.1 mg), a multi-channel parallel extraction module (including ultrasonic-assisted extraction, microwave-assisted extraction and pressurized liquid extraction and other extraction methods), an online filtration and purification module (including solid-phase extraction and membrane filtration technology) and a vacuum concentration module (with a temperature control below 40°C). The system can process 96 samples simultaneously, greatly improving the efficiency of sample pretreatment. In addition, the platform also integrates a bar code identification system and a sample management database, realizes sample full-process tracking and automatic data recording, effectively reduces human operation errors and improves data reliability. For example, in the analysis of multiple batches of ginseng samples, the platform shortens the traditional 3-day sample pretreatment work to 6 hours, and the operation consistency RSD between samples is improved to <3%.

[0052] In another alternative embodiment, an ultra-high dimensional analysis technology based on ion mobility spectrometry can be introduced. This technology adds an ion mobility spectrometry dimension to the traditional liquid chromatography-mass spectrometry system, forming an LC-IMS-MS three-dimensional analysis platform. Ion mobility spectrometry realizes the separation of gas-phase ions according to the difference in molecular collision cross section (CCS) values, adding an additional separation dimension for complex sample analysis. The system uses a high-resolution time-of-flight mass spectrometer (resolution > 40,000) and a high-performance ion mobility spectrometer (resolution > 60) to realize "retention time-drift time-m / z" three-dimensional data acquisition. This technology is particularly suitable for distinguishing isomers and structurally similar substances, and can separate components that co-elute in chromatography. For example, in the analysis of scutellarein glycosides, the 5,7,2',6'-tetrahydroxyflavone-7-O-glucoside and the 5,7,3',6'-tetrahydroxyflavone-7-O-glucoside, which are difficult to distinguish by traditional LC-MS methods, can be effectively separated and accurately identified by LC-IMS-MS technology according to their CCS value differences (210.5 Å 2 and 213.8 Å 2 ). In addition, as a characteristic parameter of molecular structure, the CCS value can be used together with retention time and accurate molecular weight to construct a more accurate compound identification database, improving the accuracy of unknown compound structure analysis.

[0053] In the embodiments of the present application, the component correlation network of a single Chinese herbal medicine is established in step S300, and the correlation network reflects the synergistic or antagonistic effects between different components, including the following steps C1-C2: C1: The component correlation data acquisition method includes: Component interaction data verified by combination therapy experiments; Statistical correlation data based on correlation analysis of component contents in multiple batches of similar samples; Component-target interaction data predicted by molecular docking and molecular dynamics simulation; Reported component synergistic effect knowledge based on literature mining.

[0054] Specifically, in C1, combination experiment is the most reliable method to obtain direct interaction between components. Experimental design usually adopts isobologram or combination index method to systematically evaluate synergistic or antagonistic effect of components in different proportions. Experimental procedures include: (1) screening the main active components in the target Chinese herbal medicine, usually selecting 5-10 representative components with high content (usually >0.1%) and clear activity (IC50 or EC50 value is clear); (2) designing the proportion gradient of component combination, usually including 9 matching points (such as component A: component B = 9:1, 8:2, 7:3,... 1:9); (3) measuring the joint activity of each matching ratio in the corresponding activity model (such as enzyme inhibition experiment, cell function experiment or animal model); (4) calculating the combination index (CI), CI <0.9 indicates synergistic effect, 0.9≤CI≤1.1 indicates additive effect, CI>1.1 indicates antagonistic effect; (5) establishing the interaction database of each pair of components, including whether there is synergistic effect, the best synergistic ratio and the synergistic strength, etc. For example, in the study of Ginkgo biloba extract, it is proved by anti-platelet aggregation experiment that flavonoids (such as quercetin, kaempferol) and terpene lactones (such as ginkgolide A, ginkgolide B) have significant synergistic effect (CI=0.72), and the best synergistic ratio is 3:2, while the same flavonoids only show additive effect (CI=0.95).

[0055] Based on the correlation analysis of multiple batches of component content is an effective method to obtain statistical correlation. This method analyzes the variation of the content of each component in different batches or samples from different sources (production area, harvesting time, processing method, etc.), and explores the potential biosynthetic correlation between components. The specific steps include: (1) collecting 30-50 batches of the same Chinese herbal medicine samples from different sources (production area, harvesting time, processing method, etc.); (2) using the verified multi-index quantitative method to determine the content of the main components (usually 15-30) in each sample; (3) calculating the Pearson correlation coefficient (r) or Spearman rank correlation coefficient (p) between components, |r| or |p|>0.8 and p<0.05 are considered to have significant correlation; (4) constructing the correlation matrix of component content, and visualizing the correlation between components by network, positive correlation indicates that they may share biosynthetic pathway or be affected by the same factors, and negative correlation indicates that they may have competitive relationship or be precursor-product relationship. For example, in the analysis of 50 batches of Huangqin samples, it is found that the contents of baicalin and wogonoside are significantly positively correlated (r=0.87, p<0.001), indicating that they may share the upstream synthesis pathway; while the content of baicalin and baicalein is significantly negatively correlated (r=-0.76, p<0.001), indicating that they may have precursor-product conversion relationship (baicalin deglycosylation to generate baicalein).

[0056] Molecular docking and molecular dynamics simulation are computational methods to predict ingredient-target interactions. This method predicts potential pharmacodynamic mechanisms and synergistic possibilities between ingredients by simulating the binding mode of ingredient molecules with protein targets. The specific process includes: (1) obtaining key target structures related to the efficacy of the target Chinese herbal medicine from protein databases (such as PDB) (such as COX-2 related to anti-inflammatory, Nrf2 related to antioxidant, etc.); (2) preprocessing the target protein, including removing water molecules, adding hydrogen, energy minimization, etc.; (3) preparing small molecule compound structures, including energy minimization, conformational search, etc.; (4) using molecular docking software (such as AutoDock Vina, Glide, etc.) to calculate the binding energy (usually in kcal / mol) and binding mode of the compound with the target; (5) molecular dynamics simulation (such as using AMBER or GROMACS software package) of the ingredient-target complex with binding energy <-7.0 kcal / mol, simulation time usually 50-100 ns; (6) analyzing the molecular dynamics trajectory, calculating the system stability (RMSD value), the interaction between key amino acid residues and small molecules (hydrogen bond, hydrophobic interaction, etc.), and the binding free energy (MM / GBSA method); (7) comparing the binding modes of different ingredients with the same target, identifying potential synergistic sites (such as two ingredients binding to different pockets of the target or different regions of the same pocket). For example, in the study of Sanqi's anti-thrombus mechanism, through molecular docking and molecular dynamics simulation, it was found that ginsenoside Rg1 (-9.2 kcal / mol) and ginsenoside Rb1 (-8.7 kcal / mol) bind to different sites of the platelet membrane glycoprotein GPIIb / IIIa complex. The former mainly interacts with the MIDAS region of the IIIa subunit, and the latter mainly interacts with the β-helix region of the IIb subunit. The combined use of the two may produce a synergistic effect of inhibiting platelet aggregation.

[0057] Literature mining is an important approach to obtain the knowledge of reported synergistic effects of ingredients. This method extracts the information of ingredient interactions from a large number of literatures through natural language processing techniques to form a structured knowledge base. The specific steps include: (1) Constructing a keyword library of Chinese herbal ingredients interaction, including terms such as "synergistic effect", "potentiation effect", "antagonistic effect", "combination effect", etc.; (2) Using literature retrieval tools (such as PubMed API, Web of Science API, etc.) to batch obtain the target Chinese herbal medicine related research literature, usually retrieving the literature resources in the past 20 years; (3) Applying natural language processing algorithms (such as named entity recognition, relationship extraction, etc.) to automatically identify the ingredient names and their interaction types described in the literature; (4) Manually reviewing the extracted information to ensure accuracy; (5) Constructing a structured ingredient interaction knowledge graph, including ingredient nodes, interaction relationships, effect strength, experimental evidence, etc. For example, through the mining of 5000 pieces of Danshen related literature, more than 200 descriptions of the interaction between tanshinones and phenolic acid ingredients were extracted, and a knowledge network containing the interaction between 15 key ingredients was formed, providing systematic literature support for the study of Danshen multi-ingredient synergistic mechanism.

[0058] C2: The construction method of ingredient association network includes: Construction of ingredient interaction network based on graph theory; Ingredient relationship prediction model based on machine learning; Using multivariate statistical analysis to identify structured relationships between ingredients; Using network pharmacology method to construct multi-level network of ingredient-target-pathway.

[0059] Specifically, in C2, the component interaction network based on graph theory is the most intuitive correlation expression. The network construction steps include: (1) taking each component in Chinese herbal medicine as a network node, and the node attributes include component name, chemical structure type, content level and other information; (2) according to the interaction data obtained in C1, a connection edge is established between the component pairs with synergistic or antagonistic relationship, and the edge attributes include interaction type (synergistic / antagonistic), action strength (such as CI value or correlation coefficient), action mechanism, etc.; (3) using network analysis software (such as Cytoscape, Gephi, etc.) for network visualization and analysis, including node degree distribution analysis, centrality analysis (such as betweenness centrality, closeness centrality, etc.) and community discovery (such as MCL algorithm, MCODE algorithm, etc.); (4) identifying the hub nodes (high degree ≥ 2 times of the average degree of the network) and key component modules (modular density ≥ 0.7) in the network, which usually represent the components or component groups that play a core role in the overall action of Chinese herbal medicine. For example, in the analysis of ginseng component correlation network, ginsenoside Rg1, Rb1 and Re were identified as hub nodes with the highest degree value and betweenness centrality, which have synergistic action with many other components; at the same time, three high-density functional modules were identified through module analysis, corresponding to immune regulation, energy metabolism and nerve protection, respectively, providing a network perspective for the research on the multi-component synergistic mechanism of ginseng.

[0060] Machine learning-based component interaction prediction model can predict the possible interactions between pairs of components that have not been experimentally verified. The model construction steps include: (1) extracting the training set from the verified component interaction data, usually including 100-300 pairs of known synergistic / antagonistic / additive component pairs; (2) calculating the structure descriptors of component molecules, including molecular fingerprints (such as MACCS, Morgan, etc.), physicochemical properties (such as molecular weight, octanol-water partition coefficient, topological polarity surface area, etc.) and 3D structure features (such as molecular shape, electrostatic potential distribution, etc.); (3) constructing feature vectors, including the structural similarity, functional similarity (such as target overlap) and physicochemical property difference of component pairs; (4) training machine learning models such as random forest (RF), support vector machine (SVM) or deep neural network (DNN), and evaluating model performance using cross-validation (usually requiring AUC>0.85); (5) applying the trained model to predict the interaction possibility and type between unknown component pairs; (6) selecting component pairs with high prediction confidence (usually probability>0.8) for experimental verification, and the verification results are fed back to optimize the model. For example, in the study of component interactions of Scutellaria, a random forest model trained based on 50 pairs of experimentally verified component interaction data (features including molecular fingerprint similarity, target action mode and pharmacokinetic properties, AUC=0.89) successfully predicted 3 pairs of previously unknown synergistic component pairs, with an experimental verification accuracy of 83%, providing an efficient way to discover new synergistic component combinations.

[0061] Multivariate statistical analysis is an effective tool to identify the structured relationships among components. The analysis procedure includes: (1) constructing a data matrix of component contents, with rows representing different samples or treatments, and columns representing different components, usually including 30-50 samples and 15-30 components; (2) pre-processing the data, including centering, standardization or log-transformation, to make components with different content levels comparable; (3) applying principal component analysis (PCA) to reduce dimensionality and visualize the relationships among components, components sharing the same principal component may have similar variation patterns; (4) using hierarchical cluster analysis (HCA) or K-means clustering to identify component groups, components clustered together may have biosynthetic or functional associations; (5) applying partial least squares regression (PLS) or canonical correlation analysis (CCA) to explore the relationships between component contents and activities, to determine the combination of components that contributes most to a specific activity; (6) using structural equation modeling (SEM) or Bayesian network analysis (BNA) to infer the causal relationships among components, including direct and indirect effects. For example, in the multi-index analysis of 30 batches of Danshen samples, PCA and HCA analysis found that tanshenones (tanshinone I, tanshinone IIA, etc.) and phenolic acids (salvianolic acid B, rosmarinic acid, etc.) formed two distinct groups, indicating that they may have different biosynthetic pathways; PLS analysis found that the content combination of tanshinone IIA and salvianolic acid B was significantly positively correlated with the anti-thrombosis activity (r 2 =0.83, p<0.001), indicating that they may be key synergistic components of Danshen's anti-thrombosis effect.

[0062] Network pharmacology method can construct multi-level network of components-targets-pathways, and comprehensively reveal the mechanism of Chinese herbal medicine. The construction process includes: (1) Collect the structure information of the identified components in the target Chinese herbal medicine, usually including 30-100 main components; (2) Use drug target prediction tools (such as SwissTargetPrediction, SEA, DRAR-CPI, etc.) or query known target databases (such as TCMSP, TCMID, BindingDB, etc.) to obtain the potential targets of the components, and screen the high-possibility targets according to the confidence (such as prediction probability >0.7 or experimentally verified targets); (3) Use target function annotation databases (such as GO, KEGG, Reactome, etc.) to analyze the biological processes and signaling pathways in which the targets are involved, and use enrichment analysis methods (such as Fisher's exact test, hypergeometric distribution test, etc.) to identify significantly enriched pathways (usually p<0.05 or FDR<0.05); (4) Collect disease-related target data (such as DisGeNET, OMIM, etc. databases), and analyze the overlap between Chinese herbal medicine targets and disease targets; (5) Construct multi-level network, including component-target layer (first layer), target-pathway layer (second layer) and pathway-disease layer (third layer), and completely display the action chain from components to drug efficacy; (6) Apply network analysis methods to identify key components (high centrality), core targets (high connectivity) and key pathways (multi-target regulation). For example, in the study of the mechanism of Coptis chinensis against diabetes, network pharmacology analysis found that alkaloid components such as berberine, palmatine and jatrorrhizine in Coptis chinensis act on 18 key targets including AMPK, PPAR-γ and GLP-1R, regulate 3 main pathway modules of glucose and lipid metabolism, insulin signal and inflammatory response, and form a multi-component-multi-target-multi-pathway synergistic regulation network, which provides a systematic explanation for the overall mechanism of Coptis chinensis against diabetes.

[0063] It should be noted that the construction of component association network is a key step to understand the overall action characteristics of Chinese herbal medicine, which breaks through the limitations of traditional single component research and reveals the synergistic action mode of "multi-component, multi-target and multi-pathway" of Chinese herbal medicine from a system level. Through the integration of experimentally verified data, statistical association analysis, computational simulation prediction and literature knowledge mining, multi-source and multi-level component association information is formed, which provides a theoretical basis for the generation of multi-component spectrum. The comprehensive application of graph theory, machine learning, multivariate statistics and network pharmacology methods not only realizes the visualization of component association network, but also provides network analysis and prediction functions, which helps to discover new synergistic component combinations and action mechanisms, and provides new ideas for the modern research and development of Chinese herbal medicine.

[0064] In an alternative embodiment, a dynamic Bayesian network (DBN) can be used to construct the ingredient association model with time-dependent relationships. This method infers the causal relationships and transformation pathways among ingredients by analyzing the dynamic changes of ingredient contents at different processing stages (e.g., processing) or in vivo metabolism. The specific implementation steps include: (1) design a time series experiment, such as collecting samples at different time points (e.g., 0 min, 5 min, 10 min...60 min) during processing, or plasma samples at different time points (e.g., 0 h, 0.5 h, 1 h...24 h) in in vivo metabolism studies; (2) determine the content change curve of the target ingredients in the samples at each time point; (3) construct a Bayesian network structure learning model, and use the structure expectation maximization (SEM) algorithm or Bayesian structure search algorithm to infer the time-dependent relationships; (4) determine the conditional probability distribution parameters in the network through parameter learning; (5) apply causal inference methods to verify the causal relationships between ingredients (e.g., intervention effect analysis); (6) construct a dynamic Bayesian network diagram to show the time sequence pattern of ingredient transformation and mutual regulation. For example, in the processing study of Fuzi, by analyzing the content change curve of aconitine-type ingredients at different processing time points, a dynamic Bayesian network including 15 ingredients such as aconitine, hypaconitine, and hypaconitine was established, successfully revealing the main detoxification pathway of aconitine de-benzoylated to benzoyl aconine, and then hydrolyzed to aconine, providing a theoretical basis for the optimization of Fuzi processing technology.

[0065] In another alternative embodiment, a knowledge graph-based system for integrating traditional Chinese medicine theory with modern scientific research can be introduced. This system systematically integrates traditional Chinese medicine theory (such as four qi and five flavors, meridian, monarch, minister, and assistant, etc.) with modern scientific research results (such as chemical ingredients, molecular mechanisms, etc.), and constructs a knowledge graph containing multi-level associations of "Chinese herbal medicine - nature - efficacy - ingredient - target - pathway". The system analyzes the medicinal nature theory recorded in ancient herbal classics (such as "Shennong Herbal Classic" and "Compendium of Materia Medica") through natural language processing and knowledge reasoning technology, and establishes a mapping relationship with modern chemical ingredient research results. For example, the "bitter cold" nature of Chinese herbal medicine is associated with specific categories of ingredients (such as alkaloids, bitter substances), and the "blood circulation" efficacy is associated with specific target pathways (such as platelet activation, vascular endothelial function regulation, etc.). This integration not only provides an explanation of the traditional theory perspective for the ingredient association network, but also verifies the scientific connotation of traditional theory through modern scientific methods, promoting the innovative development of traditional Chinese medicine theory. For example, in the study of Angelica sinensis, the system successfully associated the "warm and tonic blood vessels" efficacy in traditional theory with the molecular mechanism of ferulic acid promoting hematopoietic stem cell proliferation found in modern research, providing a modern scientific basis for traditional efficacy theory.

[0066] In the embodiments of the present application, a multi-component profile of a single Chinese herbal medicine is generated in step S400, which shows the component composition, content distribution and mutual relationship, including the following steps D1-D4: D1: The method further comprises: constructing a multi-component comprehensive feature library of a single Chinese herbal medicine, recording the physicochemical properties, structural information and biological activity of each component; based on the comprehensive feature library, performing structure speculation and function prediction on unknown components.

[0067] Specifically, in D1, the construction of the comprehensive feature library is the basic data support for the multi-component profile. The feature library contains the following core contents: (1) component basic information, such as English name, Chinese name, CAS number, molecular formula, molecular weight, structural category, etc.; (2) physicochemical property data, including melting point, boiling point, water solubility, n-octanol-water partition coefficient (LogP), hydrogen bond donor / acceptor number, polar surface area (PSA), dissociation constant (pKa), etc., which can be obtained by experimental determination or computational chemistry software (such as ChemAxon, ACD / Labs) prediction; (3) structural information, including 2D structure (SMILES, InChI code), 3D conformation (PDB format or MOL2 format), molecular fingerprint (such as MACCS, Morgan, ECFP, etc.), topological and geometric descriptors, etc.; (4) spectroscopic characteristics, including liquid chromatography-mass spectrometry data (retention time, quasi-molecular ion m / z value, main fragment ions and their relative abundance), UV maximum absorption wavelength, IR characteristic peak, NMR chemical shift value, etc.; (5) biological activity data, including in vitro activity (such as IC50, EC50, Ki value, etc.), in vivo efficacy (such as effective dose, administration method, animal model, etc.), cytotoxicity (such as CC50 value), pharmacokinetic parameters (such as Cmax, AUC, t1 / 2, etc.); (6) target information, including verified targets (directly acting targets proved by experiments) and predicted targets (potential targets predicted by computational methods), action types (such as agonists, antagonists, inhibitors, etc.) and binding modes (such as binding sites, key interactions, etc.); (7) literature citation, recording the source literature or database of each item of data to ensure data traceability.

[0068] The actual construction of the feature library is implemented by using a relational database (such as MySQL or PostgreSQL) or a NoSQL database (such as MongoDB), a primary index is established according to the component ID, and various types of feature data are stored in the form of tables or collections. The data acquisition approaches include: (1) experimental determination, obtaining the physicochemical properties, spectral characteristics, and activity data of specific components through laboratory tests; (2) literature mining, extracting component data from published literature through automatic text mining or manual sorting; (3) database integration, importing relevant data in batches from public databases (such as PubChem, ChEMBL, TCMSP, etc.); (4) computational prediction, predicting missing physicochemical properties or biological activity data using computational chemistry and machine learning methods. For example, the ginseng feature library contains about 120 identified components, including saponins, polysaccharides, volatile oils, etc., each component record contains about 50-100 different feature data items, forming more than 10,000 structured data records.

[0069] The unknown component structure inference based on the comprehensive feature library mainly relies on mass spectrometry data analysis and spectral library matching. The specific process includes: (1) extracting MS and MS / MS data of unknown components, including accurate mass of quasi-molecular ion and main fragment ions; (2) calculating possible molecular formula according to the accurate molecular weight, usually controlling the error range within 5ppm, and considering the matching degree of isotope distribution pattern; (3) comparing with the mass spectrometry data of known components in the feature library, calculating the spectral similarity (such as cosine similarity or dot product similarity), and judging whether it is an isomer of known components; (4) analyzing the fragment ion pattern to determine the possible parent nucleus structure and substituent group characteristics; (5) constructing possible candidate structures, and confirming the structural details step by step through MSn data; (6) using chemical software (such as MassFrontier, CFM-ID, etc.) to predict the theoretical mass spectrum of the candidate structure, and comparing with the measured spectrum to verify; (7) further confirming the structure by integrating the auxiliary information such as LC retention behavior and UV spectrum. For example, in the analysis of scutellaria components, the accurate molecular weight of an unknown component is m / z 431.0982 [M-H]-, the most possible molecular formula is C21H20O10 (error 2.3ppm) calculated, and the MS / MS fragment ions include m / z 311.0561 (loss of 120Da, corresponding to C-glucoside bond cleavage) and m / z 269.0457 (further loss of 42Da, corresponding to demethylation). By comparing with the fragment pattern of flavonoid components in the feature library, it is inferred that the component is 5,7,4'-trihydroxy-6-methoxy flavone-7-O-glucoside, and the inference result is further verified by the regularity of flavonoid glycoside components (polarity and sugar group number are positively correlated) through retention time.

[0070] Function prediction is mainly based on the principle of structural similarity and machine learning methods. The prediction process includes: (1) calculating the structural similarity between unknown components and known active ingredients in the feature library, such as comparing molecular fingerprint similarity using Tanimoto coefficient, or comparing conformational similarity using 3D structure superposition; (2) according to the similarity threshold (usually Tanimoto coefficient > 0.7) to infer the possible similar activity; (3) using structure-based activity prediction models (such as QSAR model, random forest model or deep neural network model) to predict the potential activity of unknown components; (4) using target prediction tools (such as SwissTargetPrediction, SEA, etc.) to predict the possible targets of unknown components; (5) based on the predicted target, perform pathway enrichment analysis to infer the possible biological processes and mechanisms of action; (6) evaluate the reliability of the prediction results, including model applicability domain analysis, prediction confidence calculation, etc.; (7) design experiments to verify the high confidence prediction results (such as prediction probability > 0.8). For example, in the study of Schisandra components, a new lignan component was found, its structure was similar to the known active ingredient Schisandrin B with a similarity of 0.82 (Tanimoto coefficient), and its P-glycoprotein inhibitory activity IC50 was predicted to be about 35 μM (prediction confidence 0.87) by random forest QSAR model. Further verified by Caco-2 cell model, it was confirmed that it indeed had moderate P-glycoprotein inhibitory activity (measured IC50 was 42 μM), indicating that the prediction had good accuracy.

[0071] D2: The method further comprises: Analyzing the variation rules of single Chinese herbal medicine multi-components under different production areas, different harvesting times and different processing technology conditions; establishing a component difference evaluation system to determine the quality grading standards of Chinese herbal medicines.

[0072] Specifically, in D2, the component variation pattern analysis is the key to understand the quality formation mechanism of Chinese herbal medicines. The analysis methods include: (1) design multi-factor experiments, systematically collect samples from different producing areas (usually including the genuine producing area and 3-5 non-genuine producing areas), different harvesting times (such as spring, summer, autumn, winter or specific solar terms), and different processing techniques (such as raw, stir-frying, baking, steaming, etc.), ensuring at least 5-10 parallel samples for each condition; (2) using the established multi-component quantitative method, determine the content of 15-30 main components in each sample; (3) applying multivariate statistical analysis methods, such as principal component analysis (PCA), partial least squares discriminant analysis (PLS-DA), orthogonal partial least squares discriminant analysis (OPLS-DA), etc., to identify the component characteristics and difference patterns under different conditions; (4) calculate the coefficient of variation (CV) and fold change (FC) under each condition, determine the most significant component variation (usually FC>2.0 and p<0.05); (5) conduct correlation analysis combined with climate data, soil data or processing parameters to reveal the key environmental factors or process parameters affecting component variation; (6) establish a component variation prediction model, such as using multiple linear regression (MLR), support vector regression (SVR), or artificial neural network (ANN) to predict the component content variation under specific conditions.

[0073] For example, in the quality evaluation research of ginseng, by analyzing 120 samples from Jilin, Liaoning, and Heilongjiang, the three main producing areas and different growth years (3 years, 4 years, 5 years, 6 years), it was found that: (1) the difference in producing area mainly affects the component ratio of ginsenosides, such as the Rg1 / Rb1 ratio in Jilin ginseng (0.4-0.6) is significantly higher than that in Heilongjiang ginseng (0.2-0.3), while the total saponin content has no significant difference; (2) the growth year mainly affects the total saponin content, the total saponin content of 5-6 year old ginseng (4.5%-6.0%) is significantly higher than that of 3-4 year old ginseng (2.5%-4.0%), and the proportion of low polarity saponins such as Ro and Rd increases with the growth year; (3) through PLS-DA model analysis, it was found that the combination of the contents of Rg1, Rb1, Re, and Rd can be used as the key indicator to distinguish different producing areas and years (model prediction accuracy > 90%); (4) multiple regression analysis showed that the organic matter content in the soil was significantly positively correlated with the total saponin content of ginseng (r=0.72, p<0.01), and the average annual temperature was significantly positively correlated with the Rg1 / Rb1 ratio (r=0.68, p<0.01), providing a scientific basis for the selection of high-quality ginseng producing areas.

[0074] In the processing technology research, taking Huangqi as an example, by comparing the component changes of four processing methods of raw Huangqi, wine Huangqi, fried Huangqi and baked Huangqi, it was found that: (1) The contents of baicalin and wogonoside in wine Huangqi were reduced by 15% and 20% respectively compared with raw Huangqi, while the contents of oquai and wogonin were increased by 40% and 35% respectively, indicating that the processing process promoted the hydrolysis of glycosides; (2) High-temperature processing (fried Huangqi and baked Huangqi) led to a significant decrease (30%-40%) in total flavonoids content, but the relative proportion of characteristic component oquai was increased; (3) OPLS-DA analysis showed that samples of different processing methods formed obviously separated clusters in the PCA score plot, indicating that processing methods had significant influence on the composition; (4) Through response surface analysis, it was found that processing temperature and time were the key factors affecting the composition transformation, among which temperature (150-180℃) had the most significant effect on glycoside hydrolysis (p<0.001), which provided a quantitative basis for the optimization of processing technology.

[0075] The component difference evaluation system is the basis for realizing the quality grading of Chinese herbal medicines. The establishment steps include: (1) Determine the evaluation index system, which usually includes the content of main active ingredients, the content of characteristic ingredients, the ratio of ingredients, and the content of total effective ingredients, etc. Each Chinese herbal medicine usually selects 3-5 key indicators, such as the evaluation system of ginseng which includes the content of total saponins, the ratio of Rg1 / Rb1 and the content of rare saponins, etc.; (2) Establish the index weight system, determine the weight coefficient of each index by analytic hierarchy process (AHP) or principal component analysis (PCA), such as giving higher weight to the component index with the strongest correlation with efficacy; (3) Set the grading threshold, determine the threshold range of each grade based on large sample statistical analysis (usually >100 samples), such as setting the top 10% of sample index range as the threshold of special grade, 10%-30% as the threshold of first grade, and 30%-70% as the threshold of second grade; (4) Establish a comprehensive scoring model, standardize each index value, and calculate the comprehensive score according to the weight, usually using 100-point system; (5) Establish the similarity evaluation method of quantitative fingerprint, combine the qualitative fingerprint characteristics and quantitative content index for comprehensive evaluation; (6) Develop evaluation software system to realize the automation and standardization of evaluation process; (7) Perform methodology validation, including repeatability evaluation (variation of scores for the same sample), reproducibility evaluation (consistency of scores in different laboratories) and discrimination accuracy (correct rate of blind sample grading).

[0076] For example, in the study of quality grading standards for Scutellaria baicalensis Georgi, an evaluation system based on the content of main components was established: (1) the contents of baicalin, wogonoside, baicalein, and wogonin were determined as basic evaluation indexes; (2) the weights of the four components were determined as 0.35, 0.25, 0.25, and 0.15, respectively, through correlation analysis with anti-inflammatory activity; (3) based on statistical analysis of 200 samples of S. baicalensis, the grading standards were set as follows: special grade (total flavonoids ≥ 18%, baicalin ≥ 10%, wogonoside ≥ 2%), first grade (total flavonoids ≥ 15%, baicalin ≥ 8%, wogonoside ≥ 1.5%), second grade (total flavonoids ≥ 12%, baicalin ≥ 6%, wogonoside ≥ 1.0%); (4) a comprehensive scoring formula was established: = 35 × (baicalin content / 12%) + 25 × (wogonoside content / 2.5%) + 25 × (baicalein content / 0.5%) + 15 × (wogonin content / 0.3%), and the total score ≥ 90 was special grade, 75-90 was first grade, 60-75 was second grade, and < 60 was unqualified; (5) similarity evaluation based on chemical fingerprint was established, and the average spectrum of special grade was used as a reference, the similarity ≥ 0.95 was excellent, 0.90-0.95 was good, 0.85-0.90 was qualified, and < 0.85 was unqualified; (6) automatic grading software based on pattern recognition was developed to realize automatic judgment from chromatographic data to quality grade, and the accuracy rate was more than 95%. This systematic quality grading system provides a scientific basis for the commercial transaction, clinical medication, and industrial quality control of S. baicalensis.

[0077] D3: The method further comprises: Based on the multi-component chromatogram of a single Chinese herbal medicine, a correlation prediction model of component-target-pathway-effect is established; A visual network of the mechanism of action of Chinese herbal medicine is generated to clarify the overall efficacy of multi-component synergistic action.

[0078] Specifically, in D3, the association prediction model is a key tool to reveal the mechanism of Chinese herbal medicine. The construction steps include: (1) ingredient-target prediction: for the 30-100 main ingredients identified in Chinese herbal medicine, the possible targets of the ingredients are predicted by target prediction methods (such as molecular docking, similarity search, machine learning model, etc.), or the known ingredient-target interaction information is collected from public databases (such as BindingDB, ChEMBL, DrugBank, etc.). The prediction parameters are usually set to high specificity (such as molecular docking score <-7.0 kcal / mol or similarity drug target overlap >60%), to ensure the reliability of the predicted targets; (2) target prediction verification: select key targets for experimental verification, such as using in vitro binding experiments (such as surface plasmon resonance, isothermal titration calorimetry, etc.) or functional experiments (such as enzyme activity inhibition, receptor activation, etc.) to confirm the accuracy of the prediction. Usually, the accuracy of the prediction is required to be >50% to ensure the effectiveness of the subsequent analysis; (3) target pathway mapping: the verified targets are mapped to biological pathway databases (such as KEGG, Reactome, WikiPathway, etc.) through bioinformatics tools (such as ClueGO, Metascape, etc.) for pathway enrichment analysis. The significance threshold is usually set to p<0.05 and FDR<0.1, and the significant pathways with enrichment factor >2.0 are selected; (4) pathway- efficacy association analysis: using medical literature knowledge base and disease target database (such as CTD, DisGeNET, etc.), the association between Chinese herbal medicine pathways and therapeutic efficacy is analyzed. Usually, hypergeometric distribution test or Fisher's exact test is used to evaluate the association significance (p<0.01); (5) model construction: based on the above analysis, an association prediction model containing four layers of ingredient-target-pathway-efficacy structure is constructed, and network inference algorithms (such as random walk, propagation coefficient, etc.) are used to quantify the association strength between each layer; (6) model verification: the accuracy of the model prediction is verified by known pharmacodynamic verification model, or the practicability of the model is verified by clinical application feedback. Usually, the prediction accuracy of the model for the main efficacy is required to be >70%; (7) new efficacy prediction: based on the verified model, the potential new efficacy or new indication of Chinese herbal medicine is predicted, to provide a theoretical basis for the expansion of clinical application.

[0079] For example, in the study of multi-component mechanism of Danshen, a correlation prediction model based on network pharmacology was constructed: (1) The structure information of 42 main components in Danshen (including tanshinones, phenolic acids and flavonoids) was collected. The target points were predicted by SwissTargetPrediction and molecular docking method, combined with the known target points reported in the literature, and a total of 189 high-confidence target points were obtained; (2) SPR experiments verified 10 pairs of key component-target interactions, such as tanshinone IIA and PPAR-γ (Kd=0.76 μM), salvianolic acid B and ACE (IC50=26 μM), which verified the reliability of the prediction; (3) Target enrichment analysis found 6 significant pathway modules, including vascular regulation (p=2.1e-8), lipid metabolism (p=4.3e-7), antioxidant stress (p=1.5e-6), anti-platelet activation (p=7.8e-6), anti-myocardial ischemia (p=2.2e-5) and anti-inflammatory (p=8.9e-5); (4) By comparing with the cardiovascular disease target database, it was found that the target points of Danshen had significant overlap with the target points of diseases such as coronary heart disease, hypertension and stroke (Jaccard index>0.4, p<0.001); (5) A four-layer network model based on random walk algorithm was constructed to quantify the component-effect correlation, and it was found that the efficacy contribution of tanshinone IIA, salvianolic acid B and cryptotanshinone was the highest (0.15, 0.12 and 0.08, respectively); (6) The prediction accuracy of the model for the main effects of Danshen (activating blood circulation to dissipate blood stasis, expanding coronary artery and anti-myocardial ischemia) reached 85%, which was verified by clinical drug feedback; (7) The model predicted that Danshen may have potential efficacy on metabolic syndrome (prediction confidence 0.73), and animal experiments preliminarily verified its effect on improving insulin resistance, providing a new idea for the clinical application of Danshen.

[0080] Mechanism visualization network is an effective tool to intuitively demonstrate the overall efficacy of Chinese herbal medicine. The construction steps include: (1) network design: determine the network hierarchy, usually including the component layer (representing the main active ingredients in Chinese herbal medicine), the target layer (representing the biological targets of ingredient action), the pathway layer (representing the biological processes in which the target participates) and the efficacy layer (representing the final therapeutic effect); (2) node attribute setting: set the visual attributes of nodes such as shape, size and color according to the characteristics of the data, such as using different shapes to represent different types of nodes (circles represent ingredients, diamonds represent targets, and squares represent pathways), using node size to represent importance (such as ingredient content, target action intensity, etc.), using node color to represent classification (such as ingredient chemical class, target function class, etc.); (3) edge attribute setting: set the style, thickness and color of the connecting edge according to the association properties, such as using edge thickness to represent the strength of action (such as binding affinity, inhibition rate, etc.), using edge color to represent the type of action (such as activation, inhibition, etc.), using edge style to represent the level of evidence (such as experimental verification, computational prediction, etc.); (4) layout optimization: adopt layout algorithms suitable for multi-level networks, such as hierarchical layout, ability-oriented layout, ring layout, etc., to ensure that the network is clear and readable, and to minimize node overlap and edge crossing; (5) interactive function design: add interactive functions to the network, such as node filtering, path highlighting, node expansion / folding, scaling and translation, etc., to facilitate user exploration of complex networks; (6) annotation information integration: integrate detailed information of ingredients, targets and pathways into the network to facilitate user query and understanding; (7) visual optimization: adjust the overall visual effect of the network, including color scheme, label layout, legend design, etc., to improve readability and aesthetics.

[0081] For example, in the visualization of the mechanism of action of Coptis, a multi-level network graph was constructed: (1) The network contains 4 levels: component layer (22 alkaloid components), target layer (78 validated targets), pathway layer (25 key pathways), and efficacy layer (6 major therapeutic effects); (2) Node settings: component nodes are set in proportion to the content (maximum content berberine as the largest node), and the color is divided into 4 categories according to the structure type (eagle class, palm leaf fangji alkaloid class, original berberine class, and tetrahydroberberine class); target node size is set according to the degree centrality, and the color is divided into 5 categories according to the function (receptor class, enzyme class, transcription factor class, ion channel class, and others); pathway node size is set according to the target enrichment degree, and the color is divided into 7 categories according to the pathway category; efficacy node size is set according to the clinical application frequency; (3) Edge settings: component-target edge thickness is set according to the binding affinity, and the color distinguishes activation (green) and inhibition (red); target-pathway edge thickness is set according to the importance of the target in the pathway; pathway-efficacy edge thickness is set according to the contribution degree; (4) Adopting a multi-level radial layout based on clustering, related nodes are clustered to reduce visual confusion; (5) Adding interactive functions to allow users to filter the network by component type, target function, or pathway category, and highlight the action path of a specific component; (6) Integrating detailed annotation information including chemical structure, mechanism of action, experimental evidence, etc., which can be viewed by clicking on the node; (7) Using a soft color scheme to ensure that key information is prominent and the overall harmony. The final visualization network intuitively shows that Coptis, through its main components such as berberine and palmatine, acts on multiple targets (such as AMPK, PPAR-γ, NF-κB, etc.), regulates multiple pathways such as glucose and lipid metabolism, inflammatory response, and antibacterial defense, and plays a whole mechanism of action of clearing heat and drying dampness, purging fire and detoxification, etc., providing a scientific basis for the modern interpretation of traditional Chinese medicine theory.

[0082] D4: The multi-component atlas of a single Chinese herb has a multi-level structure, including: a basic layer showing all detected components and their basic information; an association layer showing the interaction between components; a function layer showing the correspondence between components and biological functions; and an application layer providing data support for clinical application and new drug development of Chinese herbs.

[0083] Specifically, in D4, the construction of multi-level structure makes the multi-component atlas have rich information connotation and wide application value. The specific content and construction method of each level are as follows: The base layer is the data cornerstone of the multi-component atlas, showing all detected components in Chinese herbal medicines and their basic characteristics. The construction content includes: (1) component list: a complete list of all detected components in Chinese herbal medicines, usually including 30-200 components, each component includes basic identification information such as name, CAS number, molecular formula, etc.; (2) chemical classification: classify components according to chemical structure characteristics, such as dammarane-type triterpenoid saponins, oleanolic acid-type triterpenoid saponins, volatile oils, etc. in ginseng, forming a structured component pedigree tree; (3) content distribution: show the content level of each component in the form of quantitative data, usually expressed in mg / g or percentage, to facilitate the comparison of the abundance of different components; (4) physicochemical properties: record the key physical and chemical parameters of the components, such as polarity (logP value), molecular weight, solubility, etc. These parameters are related to the absorption and distribution characteristics of the components; (5) structure information: provide 2D structure formula and 3D conformation information of the components, which is convenient for structure comparison and structure-activity relationship analysis; (6) spectroscopic characteristics: record the chromatographic retention behavior (retention time, retention index, etc.) and mass spectrometric characteristics (molecular ion, characteristic fragment ions, etc.) of the components, which is convenient for component identification and detection method development; (7) source and distribution: mark the distribution and content difference of components in different parts of plants (roots, stems, leaves, flowers, fruits, etc.), reflecting the characteristics of component biosynthesis and accumulation. The base layer is usually presented in the form of tables, lists or databases to ensure the completeness and accuracy of the data, and is the basis for subsequent advanced analysis.

[0084] The association layer reveals the interaction between components, which is the key to understanding the overall characteristics of Chinese herbal medicines. The construction content includes: (1) structural association: based on chemical structure similarity or biosynthetic pathway, showing the structural association between components, such as precursor-product relationship, homolog relationship, etc.; (2) content correlation: based on statistical analysis of multiple batches of samples, showing the correlation between component contents, such as positive correlation (co-change) or negative correlation (mutual elimination); (3) functional synergy: based on experimental data or model prediction, showing the synergistic or antagonistic relationship between components, including synergistic mechanism, optimal ratio, etc. information; (4) metabolic transformation: showing the transformation relationship of components in the in vivo metabolic process, such as hydrolysis, oxidation, reduction, combination reaction, etc.; (5) stability influence: showing the mutual influence between components in the storage or processing process, such as some components can stabilize or accelerate the degradation of other components; (6) dissolution influence: showing the influence of components on the dissolution rate of each other, such as some components can promote the dissolution of other difficult-to-dissolve components; (7) compatibility of compound: when combined with other Chinese herbal medicines, it shows the possible interaction between components, providing a theoretical basis for compound compatibility. The association layer is usually presented in the form of network diagram, relationship matrix or heat map, which intuitively shows the complex mutual relationship between components, and provides scientific basis for understanding the integrity and synergistic effect of Chinese herbal medicines.

[0085] The correspondence between functional layer components and biological functions is the core of elucidating the material basis of Chinese herbal medicine efficacy. The construction content includes: (1) Activity spectrum: display the biological activity spectrum of each component, including in vitro activity (such as enzyme inhibition, receptor regulation, etc.) and in vivo activity (such as drug efficacy in animal models, etc.), and quantitatively compare the activity intensity and range of different components; (2) Target network: display the action relationship between components and biological targets, including action strength (such as IC50, Ki value, etc.), action type (such as agonism, antagonism, etc.) and action mode (such as binding site, conformational change, etc.); (3) Pathway regulation: display the key signal pathways and molecular networks regulated by components through targets, and explain the molecular mechanism of their efficacy, usually based on experimental verification and bioinformatics analysis; (4) Efficacy contribution: quantitatively evaluate the contribution of each component or component combination to the specific efficacy of Chinese herbal medicine, and identify the dominant components and synergistic component groups; (5) Cell impact: display the impact of components on specific cell types, such as the regulatory effect on immune cells, the inhibitory effect on tumor cells, etc.; (6) Tissue distribution: display the distribution characteristics of components in different tissues and organs in vivo, which is related to their targeted action; (7) Time sequence dynamics: display the time dynamic characteristics of component action, such as onset time, duration of action, etc. The functional layer is usually presented in the form of functional annotation network, heat map matrix or radar chart, systematically displaying the functional diversity and action mechanism of Chinese herbal medicine components, and providing scientific basis for efficacy evaluation and quality control.

[0086] The application layer is based on actual application needs, providing decision support for clinical application and new drug research and development of Chinese herbal medicine. The construction content includes: (1) Quality standard: based on multi-component data, establish a multi-index system for quality control of Chinese herbal medicine, including detection method, limit requirement and judgment standard, etc., to guide product quality evaluation; (2) Clinical guidance: based on the pharmacodynamic and pharmacokinetic characteristics of components, provide suggestions on drug dosage, administration method, administration time, etc., to optimize treatment plan; (3) Compatibility reference: based on the synergistic relationship of components, provide theoretical basis for compatibility of Chinese herbal medicine, explain the compatibility principle of classical prescriptions, and guide the development of new compound prescriptions; (4) Processing optimization: based on the transformation rule of components, guide the optimization of processing technology of Chinese herbal medicine, realize the directional regulation of specific efficacy; (5) New drug idea: based on active ingredients and action mechanism, provide lead compounds and structure modification ideas for new drug research and development, promote innovative drug discovery; (6) Safety warning: based on toxic and side component data, establish a safety risk assessment system, provide suggestions for drug safety monitoring and risk control; (7) Industry guidance: based on the formation rule of components, guide the standardized planting, processing and storage of Chinese herbal medicine, improve the quality stability of products. The application layer is usually presented in the form of decision support system, expert system or application software, which transforms basic research achievements into practical application tools, promotes the modernization development and rational application of Chinese herbal medicine.

[0087] It should be noted that the multi-level structure of the multi-component atlas breaks through the limitations of traditional single-level research, builds a complete knowledge system from component identification, relationship analysis, function explanation to application transformation, and comprehensively reflects the complex composition, internal correlation, mechanism and application value of Chinese herbal medicine. Through hierarchical data organization and visual presentation, the complex information of Chinese herbal medicine components is more systematic and structured, facilitating the understanding and application of users with different backgrounds. The basic layer provides comprehensive and accurate component data for researchers; the correlation layer reveals the interaction between components and explains the scientific connotation of the overall nature of traditional Chinese medicine; the function layer explains the pharmacodynamic material basis of Chinese herbal medicine and provides a basis for quality evaluation; the application layer directly serves clinical practice and industrial development, promoting the transformation and application of research results. This multi-level atlas structure realizes the scientific progress from "knowing the phenomenon" to "knowing the reason" and "knowing the necessity", and provides a new idea and method for modernization research of traditional Chinese medicine.

[0088] In summary, the single Chinese herbal medicine multi-component atlas construction method provided by the present application determines the target component category, uses multiple detection techniques to obtain component information, establishes a component correlation network, and generates a multi-level structure multi-component atlas, thereby constructing a systematic and comprehensive component research technical system for Chinese herbal medicine. This method breaks through the limitations of traditional single component research, realizes the overall grasp of the complex composition of Chinese herbal medicine, reveals the synergistic relationship between components, provides a scientific basis for traditional Chinese medicine quality control, efficacy evaluation and new drug research, and promotes the in-depth development of modernization research of traditional Chinese medicine.

[0089] Example 3 The above is a schematic scheme of a single Chinese herbal medicine multi-component atlas construction method. It should be noted that the technical scheme of the single Chinese herbal medicine multi-component atlas construction system is the same as the technical scheme of the single Chinese herbal medicine multi-component atlas construction method described above, and the technical scheme of the single Chinese herbal medicine multi-component atlas construction system in this embodiment is not described in detail. The details can be referred to the description of the technical scheme of the single Chinese herbal medicine multi-component atlas construction method.

[0090] The embodiment also provides a single Chinese herbal medicine multi-component atlas construction system, comprising: A component category determination module is configured to determine the target component category of the single Chinese herbal medicine, wherein the target component category includes active components, marker components, auxiliary components and toxic and side components. A component information acquisition module is configured to acquire qualitative and quantitative information of each type of target component in the single Chinese herbal medicine by using multiple detection techniques. An association network establishment module is configured to establish a component association network of the single Chinese herbal medicine, wherein the association network reflects the synergistic or antagonistic effect between different components. The multi-component atlas generation module is configured to generate a multi-component atlas of a single Chinese herbal medicine, and the atlas displays component composition, content distribution and mutual relationship.

[0091] The embodiment also provides an electronic device suitable for the case of constructing a multi-component atlas of a single Chinese herbal medicine, which comprises a memory and a processor.

[0092] The embodiment also provides a storage medium having a computer program stored thereon, and the computer program is executed by a processor to implement the method for constructing a multi-component atlas of a single Chinese herbal medicine.

[0093] The storage medium proposed by the embodiment and the method for constructing a multi-component atlas of a single Chinese herbal medicine proposed by the above embodiment belong to the same inventive concept, and the technical details not described in the embodiment can be referred to the above embodiment, and the embodiment and the above embodiment have the same beneficial effects.

[0094] Through the above description of the embodiments, those skilled in the art can clearly understand that the present application can be realized by means of software and necessary general hardware, and of course can also be realized by hardware. Based on such understanding, the technical solutions of the present application can be embodied in the form of a software product, and the computer software product can be stored in a computer readable storage medium, such as a floppy disk, a ROM, a RAM, a FLASH, a hard disk or an optical disk, etc., and includes a number of instructions to make a computer device (which can be a personal computer, a server, or a network device, etc.) execute the methods of various embodiments of the present application.

[0095] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present application but not limit the present application, and although the present application has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present application can be modified or replaced by equivalents without departing from the spirit and scope of the present application, and all of them should be covered in the scope of the claims of the present application.

Claims

1. A method for constructing multi-component maps of a single traditional Chinese medicine, characterized by: include, Identify the target component category of a single traditional Chinese medicine, wherein the target component category includes active ingredients, marker components, auxiliary components, and toxic and side effects components; Multiple detection technologies were employed to obtain qualitative and quantitative information on various target components in the single Chinese herbal medicine. Establish a component association network for the single Chinese herbal medicine, wherein the association network reflects the synergistic or antagonistic effects between different components; A multi-component spectrum of the single Chinese herbal medicine is generated, which shows the component composition, content distribution and interrelationships.

2. The method for constructing a multi-component map of a single traditional Chinese medicine as described in claim 1, characterized in that: The target component categories also include: chemical fingerprint characteristic components, functionally dominant components, and structurally similar component groups.

3. The method for constructing a multi-component map of a single traditional Chinese medicine as described in claim 2, characterized in that: The various detection technologies include: A combination of chromatographic separation technology, mass spectrometry identification technology, spectral analysis technology and biological detection technology.

4. The method for constructing a multi-component map of a single traditional Chinese medicine as described in claim 3, characterized in that: The method further includes: Construct a comprehensive feature library of multiple components of a single Chinese herbal medicine to record the physicochemical properties, structural information and biological activities of each component; Based on the comprehensive feature library, structural inference and functional prediction are performed on unknown components.

5. The method for constructing a multi-component map of a single traditional Chinese medicine as described in claim 4, characterized in that: The method further includes: Analyze the variation patterns of multiple components in a single Chinese herbal medicine under different origins, harvest times, and processing techniques. Establish a component difference evaluation system and determine the quality grading standards for traditional Chinese medicine.

6. The method for constructing a multi-component map of a single traditional Chinese medicine as described in claim 5, characterized in that: The method further includes: Based on the multi-component spectrum of the single Chinese herbal medicine, a correlation prediction model of component-target-pathway-efficacy is established. A visualization network of the mechanism of action of traditional Chinese medicine was generated to elucidate the overall efficacy of the synergistic effect of multiple components.

7. The method for constructing a multi-component map of a single traditional Chinese medicine as described in claim 6, characterized in that: The multi-component spectrum of the single traditional Chinese medicine has a multi-level structure, including: The base layer displays all detected components and their basic information; The relationship layer displays the interactions between components; The functional layer shows the correspondence between components and biological functions; The application layer provides data support for the clinical application of traditional Chinese medicine and the research and development of new drugs.

8. A system for constructing multi-component maps of a single traditional Chinese medicine, based on the method for constructing multi-component maps of a single traditional Chinese medicine as described in any one of claims 1 to 7, characterized in that: It also includes a component category determination module for determining the target component category of a single Chinese herbal medicine, wherein the target component category includes active ingredients, marker components, auxiliary components, and toxic and side effects components; The component information acquisition module is used to acquire qualitative and quantitative information of various target components in the single Chinese herbal medicine using multiple detection technologies; The association network establishment module is used to establish an association network of the components of the single Chinese herbal medicine, wherein the association network reflects the synergistic or antagonistic effects between different components; A multi-component spectrum generation module is used to generate a multi-component spectrum of the single Chinese herbal medicine, which displays the component composition, content distribution and interrelationships.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, it implements the steps of the method for constructing a multi-component map of a single traditional Chinese medicine as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, it implements the steps of the method for constructing a multi-component map of a single traditional Chinese medicine as described in any one of claims 1 to 7.

Images