Analysis method for characteristic chemical components of Qinling Mountain middle-peak honey

Abstract

The invention belongs to the technical field of honey analysis, and particularly relates to a method for analyzing characteristic chemical components of Qinling Mountain medium-peak honey. The analysis method comprises the following steps: obtaining different extraction layers, and carrying out mass spectrometry on each layer of sample; constructing a molecular network by mass spectrum data through a GNPS platform, and comparing the mass spectrum data with a database to obtain primary identification; unmatched nodes predict the skeleton by using an NAP tool to obtain a secondary result; then fractions are separated, and an unknown structure is deduced by combining an SMART technology with nuclear magnetism; and finally, synthesizing the three types of results to determine the characteristic chemical components of the honey. Compared with a single chromatography-mass spectrometry technology, the analysis method provided by the invention not only realizes data visualization, but also reduces the dependence of chemical component identification on a database to a certain extent, and provides reference for rapid characterization of chemical components in a honey crude extract or fraction.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of honey analysis, and particularly relates to a method for analyzing characteristic chemical components of Qinling Zhongfeng honey. BACKGROUND

[0002] The difference in plant species of nectar sources can significantly affect the nutritional components and functional properties of honey, and such difference can also lead to fraudulent behaviors in plant sources or geographical origins and uneven quality in the honey market. Therefore, it is of great significance to establish a precise traceability identification system to protect the quality of honey and the rights and interests of consumers. In the identification research of honey plant sources, metabolomics technology has attracted much attention due to its comprehensive analysis advantage: this technology can systematically analyze the composition of metabolites in honey, not only identify biomarkers with species specificity, but also deeply reveal the metabolic differences between different single-flower nectar sources, thereby providing a scientific basis for constructing a high-precision species identification model. At present, metabolomics methods based on mass spectrometry, such as liquid chromatography-mass spectrometry (LC-MS) and gas chromatography-mass spectrometry (GC-MS), dominate in this field. However, traditional chromatography-mass spectrometry technology has obvious limitations, and its analysis efficiency is highly dependent on the existing database. For chemical components not included in the database, in most cases, they can only be labeled as "unknown" and cannot obtain any information. In addition, traditional chromatography-mass spectrometry technology often ignores low-abundance and co-flowing compounds in the result division of the large amount of mass spectrometry data, and the analysis process is complicated and easy to miss information. In summary, it is difficult to accurately analyze the chemical components not included in the database and with low abundance in the process of metabolite analysis using traditional chromatography-mass spectrometry technology. SUMMARY

[0003] To solve the problem that it is difficult to accurately analyze the chemical components not included in the database and with low abundance in the process of metabolite analysis using traditional chromatography-mass spectrometry technology in the prior art, the application provides a method for analyzing characteristic chemical components of Qinling Zhongfeng honey. To achieve the above object, the application adopts the following technical solutions.

[0004] The application provides a method for analyzing characteristic chemical components of Qinling Zhongfeng honey, which comprises the following steps: Different organic solvents are used to extract Qinling Zhongfeng honey to obtain different extraction layers, and mass spectrometry data of chemical components in the different extraction layers are obtained.

[0005] Based on the GNPS platform, the mass spectrometry data of the chemical components in the different extraction layers are clustered and integrated, and visualized as a molecular network.

[0006] The molecular network is matched with a natural product database of the GNPS platform to obtain an annotation result of a matched visualization node; and a NAP annotation tool is used to annotate the visualization node in the molecular network that is not matched to predict a possible chemical skeleton and structure type.

[0007] The different extraction layers are respectively re-extracted and distilled, the obtained distillates are subjected to nuclear magnetic analysis, SMART technology is applied to infer specific structure types of chemical components in the distillates that are different from the annotation result of the visualization node and the possible chemical skeleton and structure type, and an inference result is obtained.

[0008] The annotation result of the visualization node, the possible chemical skeleton and structure type, and the inference result are subjected to confidence score, priority decision and retest verification, and finally the characteristic chemical component type of Qinlingzhongfeng honey is determined.

[0009] The analysis method of the characteristic chemical component of Qinlingzhongfeng honey provided by the application first extracts Qinlingzhongfeng honey with different organic solvents to obtain different extraction layers, obtains mass spectrum data of chemical components in the different extraction layers; based on the GNPS platform, the mass spectrum data of the chemical components in the different extraction layers are clustered and integrated, and visualized as a molecular network; the molecular network is matched with a natural product database of the GNPS platform to obtain an annotation result of a matched visualization node; a NAP annotation tool is used to annotate the visualization node in the molecular network that is not matched to predict a possible chemical skeleton and structure type; the different extraction layers are respectively re-extracted and distilled, the obtained distillates are subjected to nuclear magnetic analysis, SMART technology is applied to infer specific structure types of chemical components in the distillates that are different from the annotation result of the visualization node and the possible chemical skeleton and structure type, and an inference result is obtained: the annotation result of the visualization node, the possible chemical skeleton and structure type, and the inference result are subjected to confidence score, priority decision and retest verification, and finally the characteristic chemical component type of Qinlingzhongfeng honey is determined. The method provided by the application first uses the GNPS molecular network analysis based on LC-MS technology to systematically analyze the chemical components of each extraction layer in combination with the NAP annotation tool; and further uses the SMART method based on NMR technology to predict and verify the structure of the chemical components of the main distillates of each extraction layer. The integrated strategy effectively solves the problem that compounds not included in the database and low-abundance components are difficult to accurately analyze in the metabolite analysis process.

[0010] Preferably, the characteristic chemical component type of Qinlingzhongfeng honey includes amides, fatty acids and derivatives thereof, glycosides, alkaloids, terpenes and anthraquinones.

[0011] Preferably, the SMART technique is applied to the obtained fractions to deduce the specific structure types of the unidentified compounds therein by the filtering strategies of accurate neutral loss, characteristic fragments and mass deficit.

[0012] Preferably, the step of obtaining the molecular network is: The mass spectrum data of the chemical components in the different extraction layers are converted into data files in the.mzXML format, and the converted data files in the.mzXML format are uploaded to the GNPS platform, and the GNPS platform automatically performs the process of "spectrum de-redundancy-molecular fingerprint extraction-all-against-all similarity calculation-force-directed network layout" to complete the construction of the molecular network.

[0013] Preferably, the converted data files in the.mzXML format are imported into the Feature-Based Molecular Networking workflow of the GNPS platform, the mass deviation of the parent ions and fragment ions is set to 0.02 Da, the similarity threshold is ≥ 0.7, the minimum matching fragment ions are 6, the cluster minimum size is 2, the TopK edge number is set to 10, and the maximum connected component is set to 100, and the GNPS platform automatically completes de-redundancy, similarity calculation, TopK pruning and giant cluster segmentation, and finally outputs an interactive molecular network map.

[0014] Preferably, the different organic solvents include chloroform, ethyl acetate and n-butanol, and the obtained different extraction layers include a chloroform extraction layer, an ethyl acetate extraction layer and an n-butanol extraction layer.

[0015] Preferably, before the Qinling Apis cerana cerana honey is extracted in sequence using chloroform, ethyl acetate and n-butanol, the Qinling Apis cerana cerana honey is dispersed with a NaCl aqueous solution containing a mass fraction of 2% NaCl; The mass of the Qinling Apis cerana cerana honey to the volume of the NaCl aqueous solution is 2.25 kg: 1 L~3 L.

[0016] Preferably, the different extraction layers are respectively subjected to re-extraction and distillation, the obtained fractions are subjected to nuclear magnetic analysis, and the SMART technique is applied to deduce the specific structure types of the unidentified compounds in the obtained fractions, and the step of obtaining the deduced results is: The chloroform extraction layer is dissolved in methanol to obtain a chloroform extraction layer methanol insoluble part and a chloroform extraction layer methanol soluble part; the chloroform extraction layer methanol soluble part is separated by a chromatographic method to obtain a plurality of chloroform extraction layer methanol soluble part corresponding fractions; and the ethyl acetate extraction layer is separated by a chromatographic method to obtain a plurality of ethyl acetate extraction layer corresponding fractions.

[0017] The methanol-insoluble fraction of the chloroform extract layer, the corresponding fractions of the chloroform extract layer soluble in methanol, and the corresponding fractions of the ethyl acetate extract layer are dissolved in deuterated solvents, respectively, and the NMR data (NMR-HSQC data) are collected.

[0018] The NMR data of each fraction are subjected to automatic peak extraction, chemical shift clustering, and skeleton fragment identification by the SMART-NMR platform to generate a prediction list of the assumed chemical skeleton and / or structure type corresponding to each fraction.

[0019] The prediction list is associated with the visual nodes in the molecular network that are not annotated or have low matching degrees: if the parent ion m / z of a visual node deviates from the accurate mass of the assumed skeleton in the prediction list by ≤5 ppm, and the number of matches between the MS / MS fragments of the visual node and the main theoretical fragments of the assumed chemical skeleton is ≥4, then the visual node is annotated as the corresponding skeleton or structure type, thereby inferring the specific structure type of the uncharacterized compounds in the fraction.

[0020] Preferably, the mass spectrometry data of the chemical components in different extract layers are obtained by detecting the different extract layers using ultra-high performance liquid chromatography-quadrupole-time-of-flight tandem mass spectrometry (UPLC-Q-TOF-MS / MS).

[0021] The chromatographic conditions during detection are as follows: A Thermo Scientific™ Q Exactive Plus quadrupole-electric field orbitrap high-resolution mass spectrometry system is used; an ACQUITY UPLC BEH C18 chromatographic column with a size of 2.1×50 mm and a particle size of 1.7 µm is used; the column temperature is 30℃; the flow rate is 0.20 mL / min; and the injection volume is 3.0 μL.

[0022] The mobile phase includes mobile phase A and mobile phase B; the mobile phase A is 0.1% formic acid water (volume percent), and the mobile phase B is acetonitrile.

[0023] Preferably, the chromatographic conditions during detection further include the elution program shown in the following table:

[0024] Preferably, the mass spectrometry conditions during detection are as follows: In positive and negative ion modes, an electrospray ionization (ESI) source is equipped, the nebulizer pressure is 35 psi; the sheath gas temperature is 350℃, the flow rate is 10 AU; the auxiliary gas flow rate is 1.0 AU; the capillary temperature is 320℃; the MS scan range is 100 m / z ~1500 m / z , and the MS / MS range is 50 m / z ~1500m / z The collision energy is 20 eV, 40 eV or 60 eV.

[0025] Preferably, before extracting Qinling Apis cerana cerana honey using chloroform, ethyl acetate and n-butanol in sequence, the Qinling Apis cerana cerana honey is dispersed with a NaCl aqueous solution containing a mass fraction of 2%.

[0026] The mass of the Qinling Apis cerana cerana honey to the volume of the NaCl aqueous solution is 2.25 kg: 1 L~3 L.

[0027] Compared with the prior art, the present application has the following beneficial effects: 1. The present application provides an analysis method for characteristic chemical components of Qinling Apis cerana cerana honey. The method first extracts Qinling Apis cerana cerana honey with multiple organic solvents, and determines each layer sample by ultra-high performance liquid chromatography-quadrupole-time-of-flight tandem mass spectrometry; the mass spectrometry data are used to construct a molecular network on the GNPS platform and compared with a database to obtain a primary identification result; unmatched nodes are used to predict the skeleton by NAP tools to obtain a secondary prediction result; then the fractions are separated by column chromatography-semi-preparative liquid chromatography, and unknown structures are inferred by SMART technology combined with nuclear magnetic resonance to obtain an inference result; finally, the three types of results are aligned according to the retention time, and the confidence score, priority decision and retest are performed, and the nodes with a score of ≥0.7 are reserved to determine the type of characteristic chemical components of Qinling Apis cerana cerana honey. The method provided by the present application first adopts GNPS molecular network analysis based on LC-MS technology, combines NAP annotation tools to systematically analyze the chemical components of each extraction layer; further uses SMART method based on NMR technology to predict and verify the chemical components of the main fractions of each extraction layer. The integrated strategy effectively solves the problem that compounds not included in the database and low-abundance components are difficult to accurately analyze in the metabolite analysis process.

[0028] Specifically, the present application is based on the problem of complex and low content of phytochemical components in Qinling Apis cerana cerana honey. The GNPS molecular network is used to identify and cluster the chemical components of each extraction layer. Further combined with NAP annotation tool and SMART analysis, the chemical skeleton and structure type of the compounds that have not been successfully identified are predicted. Compared with the traditional single chromatography-mass spectrometry technology, GNPS can automatically classify and organize hundreds of thousands of mass spectrometry signals according to chemical correlation. Even if some compounds are unknown (i.e. not included in the existing database), as long as their MS / MS spectrum is similar to that of known compounds or a certain group of compounds, they will be clustered into the same "molecular family" and form a visual "chemical map". In order to deeply analyze the unknown components in the molecular network, the NAP tool is used to annotate the nodes in the network that have no spectral library matching. At the same time, with the help of SMART analysis method, the structure type of natural products in the main fraction of the extraction layer is predicted, thereby effectively supplementing and verifying the analysis results of GNPS molecular network. This integrated strategy not only realizes the visual analysis of data, but also to some extent reduces the dependence on existing chemical databases, and provides a feasible method reference for rapid characterization of chemical components in honey crude extract or its fractions.

[0029] 2、The present application uses UPLC-Q-TOF-MS / MS, GNPS platform, NAP annotation tool and SMART technology to analyze the chemical components of Qinling Apis cerana cerana honey. First, the Apis cerana cerana honey is extracted with chloroform, ethyl acetate and n-butanol in turn to obtain the mass spectrometry information of each extraction layer, and then the mass spectrometry information is integrated and visualized into a molecular network based on GNPS. Through comparison with the GNPS database, 28 metabolites are identified, including amides, fatty acids and their derivatives, polyphenols and glycosides. It is indicated that the nodes constituting the same molecular cluster have similar chemical information. Among them, amides and fatty acids and their derivatives are the main detected components. Importantly, the detection results also include a vitamin B2 metabolite named dimethyl isoalloxazine. According to Du et al (Du Y, Zhu H, Qiao J, et alThe results of the study on "Characteristic components and authenticity evaluation of Chinese honeys from three different botanical sources[J]. Journal of Agricultural and Food Chemistry, 2023, 71(45): 17284-17294" may explain the bright yellow color of the honey. For compounds that were not successfully identified, the structures were predicted using the NAP annotation tool. The results showed that some nodes were identified as alkaloids, lipids and lipid molecules, phenylpropanoids and polyketides, and organic heterocyclic compounds, indicating that compounds related to these categories of skeletons may exist in the honey. These analytical results suggest that alkaloids, lipids and lipid molecules may be the most abundant chemical components in Qinling honey, while glycosides have a greater dominance in content.

[0030] Based on the results of GNPS molecular network analysis, the methanol-soluble fraction of chloroform extract (HM-C) and ethyl acetate extract (HM-E) were fractionated by column chromatography and semi-preparative liquid chromatography, respectively. The structure types of compounds in the methanol-insoluble fraction of chloroform extract (the dichloromethane-soluble fraction of chloroform extract, HD-C) and each fraction (C1~C3, E1~E3) were inferred based on the accurate identification of small molecules. HD-C was inferred to contain a large number of oxides, such as macrocyclic diterpenes or cyclic oxides. In the fractions of HM-C, C-1 and C-2 were speculated to be clusters of terpenoids, but C-1 was more likely to contain a macrocyclic lactone structure similar to Rediocide C, and the structure types of compounds in C-2 were more similar to Jiadifenoic acid. The main components of C-3 were identified as alkaloids, and were most likely similar to indole alkaloids, represented by Kopsiyunnanine C. In the analysis results of the sub-fraction of HM-E, E-2 had the highest cosine similarity score to Normelicopicine, and there were 5 anthraquinones in the top 10 structures, suggesting that E-2 contained anthraquinones and acridinones, which were different from the alkaloid structures in C-3. This finding supplemented the annotation results of NAP annotation tool for alkaloids in HM-C and HM-E. The analysis results of E-3 showed that there might be long-chain aliphatic hydrocarbon structures, including long-chain fatty acids, esters and amides, which not only matched the results of GNPS, but also inferred that the NAP tool annotation of lipids and lipid-like molecules in HM-E was similar to Paracaseolide A and Callyspongidic acid. Finally, E-1 was inferred to be a cluster of glycosides represented by Geniposidicacid, which was similar to 6α-Hydroxygeniposide, similar to the matching results of GNPS in negative mode.

[0031] The analysis strategy combining GNPS and NAP annotation tool reduces the dependence on databases to some extent compared to traditional analysis methods. The analysis results of SMART technology not only match the GNPS database and NAP annotation results, but also provide more specific explanations for some NAP annotation results. The discovery of terpenoids (including macrocyclic lactone structures) provides key clues for the analysis of uncharacterized compounds. In summary, the SMART-based analysis not only overcomes the limitations of sample solubility in mass spectrometry-based analysis, but also makes the prediction of the structure types of compounds in total extract or fractions more concrete. This analysis method is rapid and lays a theoretical foundation for quickly identifying the structure types of metabolites in honey crude extract and fractions. BRIEF DESCRIPTION OF DRAWINGS

[0032] Figure 1 Flow chart for sample preparation in the present application.

[0033] Figure 2 Molecular network diagram in POS (A) and NEG (B) mode in the present application; wherein the numbers correspond to the compound structures and Figure 13 the compound numbers in Table 2.

[0034] Figure 3 NAP annotation result diagram in POS mode in the present application; wherein the numbers correspond to the compound structures and Figure 5 the compound numbers in Table 2.

[0035] Figure 4 NAP annotation result diagram in NEG mode in the present application; wherein the numbers correspond to the compound structures and Figure 5 the compound numbers in Table 2.

[0036] Figure 5 Structure formula diagram of NAP annotation result in the present application.

[0037] Figure 6 SMART analysis result diagram (HD-C) in the present application; wherein: A is a HSQC spectrum; B is a top 10 structure diagram of cosine similarity score.

[0038] Figure 7 SMART analysis result diagram (C-1) in the present application; wherein: A is a HSQC spectrum; B is a top 10 structure diagram of cosine similarity score.

[0039] Figure 8 SMART analysis result diagram (C-2) in the present application; wherein: A is a HSQC spectrum; B is a top 10 structure diagram of cosine similarity score.

[0040] Figure 9 SMART analysis result diagram (C-3) in the present application; wherein: A is a HSQC spectrum; B is a top 10 structure diagram of cosine similarity score.

[0041] Figure 10 SMART analysis result diagram (E-1) in the present application; wherein: A is a HSQC spectrum; B is a top 10 structure diagram of cosine similarity score.

[0042] Figure 11Figure E-2 is a SMART analysis result chart in the present application; wherein: A is a HSQC spectrum; B is a top 10 structure chart of cosine similarity score.

[0043] Figure 12 Figure E-3 is a SMART analysis result chart in the present application; wherein: A is a HSQC spectrum; B is a top 10 structure chart of cosine similarity score.

[0044] Figure 13 Figure E-4 is a structure chart of GNPS matching result in the present application. DETAILED DESCRIPTION

[0045] The present application will be described in detail below with reference to the accompanying drawings and specific examples, but should not be understood as limiting the present application. If not specifically stated, the technical means used in the following examples are conventional means well known to those skilled in the art, and the materials, reagents, etc. used in the following examples, if not specifically stated, can be obtained from commercial channels.

[0046] Example 1 I. Method The sample preparation flow chart is shown in Figure 1 .

[0047] 1. Sample pretreatment (1) 2.25 kg of Qinling mid-bee honey was weighed, dispersed with 2 L of 2% NaCl aqueous solution (w:v), and then 4 L of chloroform, ethyl acetate and n-butanol were added in sequence for liquid-liquid extraction at room temperature, and each extraction was performed for 48 h, and the extract was concentrated under reduced pressure and dried to obtain: chloroform extract layer (2.2 g), ethyl acetate extract layer (1.64 g) and n-butanol extract layer (62.42 g).

[0048] Among them, the source of Qinling mid-bee honey is in Zhashui County, Shangluo City, Shaanxi Province.

[0049] (2) The chloroform extract layer was divided into two parts according to its solubility in methanol, numbered as HM-C (methanol-soluble part of chloroform extract layer, 0.68 g) and HD-C (methanol-insoluble part of chloroform extract layer, dichloromethane-soluble, 1.36 g); the ethyl acetate and n-butanol extract layers were numbered as HM-E (ethyl acetate extract layer) and HM-B (n-butanol extract layer), respectively, and the sample preparation flow chart is shown in Figure 1 .

[0050] (3) HM-C was dissolved in chromatographic methanol, centrifuged at 14,000 rpm for 10 min, and then the supernatant was filtered through a 0.22 μm nylon filter to obtain a 1 mg / mL test solution; HM-E was dissolved in chromatographic methanol, centrifuged at 14,000 rpm for 10 min, and then the supernatant was filtered through a 0.22 μm nylon filter to obtain a 1 mg / mL test solution; HM-B was dissolved in chromatographic methanol, centrifuged at 14,000 rpm for 10 min, and then the supernatant was filtered through a 0.22 μm nylon filter to obtain a 1 mg / mL test solution.

[0051] All the above solutions were stored at 4°C until UPLC-Q-TOF-MS / MS detection.

[0052] (4) HM-E was redissolved in chromatographic methanol, filtered through a 0.22 μm filter, and then separated by semi-preparative reverse phase HPLC (semi-preparative liquid chromatograph, HPLC-NP5007C, Jiangsu Hanbang Technology Co., Ltd.) using H2O-MeOH as the mobile phase (UV detection wavelength of 210 nm and 254 nm, 5 mL / min) gradient elution to obtain three sub-components (E-1~E-3). E1 was collected by H2O-MeOH at 95:5 (elution gradient, v / v)~70:30 (elution gradient, v / v), E2 was collected by H2O-MeOH at 70:30 (elution gradient, v / v)~35:65 (elution gradient, v / v), and E3 was collected by H2O-MeOH at 35:65 (elution gradient, v / v)~0:100 (elution gradient, v / v).

[0053] HM-C was separated by column chromatography due to poor solubility (partially slightly soluble in methanol). HM-C was separated by silica gel (200-300 mesh) column chromatography eluted with dichloromethane / methanol gradient (v / v, 100:0-0:100) to obtain three fractions (C-1~C-3). C1 was collected by CH2Cl2-MeOH at 100:1 (elution gradient, v / v)~80:1 (elution gradient, v / v), C2 was collected by CH2Cl2-MeOH at 80:1 (elution gradient, v / v)~20:1 (elution gradient, v / v), and C3 was collected by CH2Cl2-MeOH at 20:1 (elution gradient, v / v)~0:100 (elution gradient, v / v). Each fraction was concentrated under reduced pressure and evaporated to dryness, and the solid residue was dissolved in 600 μL of deuterated reagent and transferred to an NMR tube for NMR detection.

[0054] 2. Sample LC-MS and NMR-HSQC data collection LC-MS detection was performed on the methanol-soluble part of the chloroform extract (HM-C), the ethyl acetate extract (HM-E) and the n-butanol extract (HM-B) of each layer of the honey of Apis cerana, and GNPS analysis data was collected; HSQC detection was performed on the dichloromethane-soluble part of the chloroform extract (HD-C, i.e. the methanol-insoluble part of the chloroform extract), the three fractions of HM-C (C-1~3) and the three fractions of HM-E (E-1~3), and SMART analysis data (NMR-HSQC data) was collected.

[0055] The instrument conditions for LC-MS detection are as follows: (1) Chromatographic conditions: Thermo Scientific™ Q Exactive Plus quadrupole-electrostatic field orbitrap high-resolution mass spectrometry system was used, and ACQUITY UPLC BEH C18 chromatographic column (2.1 mm x 50 mm, 1.7 µm, Waters) was configured to complete UPLC-Q-TOF-MS / MS analysis in positive and negative ion modes.

[0056] The column temperature was kept at 30℃, the flow rate was kept at 0.2 mL / min, and the injection volume was 3 µL.

[0057] The mobile phase consisted of 0.1% (v / v) formic acid water (A) and acetonitrile (B).

[0058] The gradient elution program is shown in Table 1.

[0059] Table 1 Elution program Wherein, it is to be noted that with the extension of the (retention) time, the volume of mobile phase B is gradually increased and then decreased to the initial concentration, and the volume of mobile phase A is gradually decreased and then increased and then decreased to the initial concentration. Wherein, at time 1 min, the volume of mobile phase B is 5%, and the volume of mobile phase A is 95%; at time 1 min~10 min, the volume of mobile phase B is gradually increased from 5% to 15%, and the volume of mobile phase A is gradually decreased from 95% to 85%; at time 10 min~14 min, the volume of mobile phase B is gradually increased from 15% to 20%, and the volume of mobile phase A is gradually decreased from 85% to 80%; at time 14 min~17 min, the volume of mobile phase B is gradually increased from 20% to 30%, and the volume of mobile phase A is gradually decreased from 80% to 70%; at time 17 min~22 min, the volume of mobile phase B is gradually increased from 30% to 40%, and the volume of mobile phase A is gradually decreased from 70% to 60%; at time 22 min~29 min, the volume of mobile phase B is gradually increased from 40% to 55%, and the volume of mobile phase A is gradually decreased from 60% to 45%; at time 29 min~34 min, the volume of mobile phase B is gradually increased from 55% to 70%, and the volume of mobile phase A is gradually decreased from 45% to 30%; at time 34 min~39 min, the volume of mobile phase B is gradually increased from 70% to 100%, and the volume of mobile phase A is gradually decreased from 30% to 0; at time 39 min~42 min, the volume of mobile phase B is 100%, and the volume of mobile phase A is 0; at time 42 min~44 min, the volume of mobile phase B is gradually decreased from 100% to the initial concentration 5%, and the volume of mobile phase A is gradually increased from 0 to the initial concentration 95%; at time 44 min~47 min, the volume of mobile phase B remains the initial concentration 5%, and the volume of mobile phase A remains the initial concentration 95%.

[0060] 0~1min, 5% (v / v) B; 1min~10min, 5% (v / v)~15% (v / v) B; 10min~14min, 15% (v / v)~20% (v / v) B; 14min~17min, 20% (v / v)~30% (v / v) B; 17min~22min, 30% (v / v)~40% (v / v) B; 22min~29min, 40% (v / v)~55% (v / v) B; 29min~34min, 55% (v / v)~70% (v / v) B; 34min~39min, 70% (v / v)~100% (v / v) B; 39min~42min, 100% (v / v) B; 42min~44min, 100 (v / v)%~5% (v / v) B; 44min~47min, 5% (v / v) B.

[0061] (2) Mass spectrometry conditions: Mass spectrometry data acquisition was performed in positive and negative ion modes, equipped with an electrospray ionization (ESI) source, with a nebulizer pressure of 35 psi; the sheath gas temperature was 350 °C, with a flow rate of 10 AU; the auxiliary gas flow rate was 1 AU; the capillary temperature was 320 °C. The mass spectrometer was operated in a data-dependent acquisition mode, with a MS scan range of 100 m / z 1500 m / z , a MS / MS range of 50 m / z 1500 m / z , and a collision energy of 20 eV / 40 eV / 60 eV.

[0062] (3) NMR-HSQC data acquisition: Bruker Avance 600 DMX nuclear magnetic resonance instrument. Among them, the conditions when HSQC detection are as follows:

[0063] Using a 600 MHz double-tapered gradient probe, using the hsqc sequence, 1H spectrum width 20 ppm, 13C spectrum width 250 ppm, t1: 256 increments, NS=8, temperature 298 K.

[0064] 3. Data processing GNPS molecular network: The mass spectrometry raw data is converted into.mzXML format using MSconvert software (ProteoWizard), and the converted data file is uploaded to the GNPS platform using Win SCP software. After completing the molecular network construction, it is exported as a Cytoscape file to visualize node connections. In the generated molecular network: Precursor Ion Mass Tolerance and Fragment Ion Mass Tolerance are both set to 0.02 Da; the similarity index between clusters is less than 0.7 and will not be considered; Minimum Cluster Size is 2 (clusters containing less than 2 similar spectra are discarded); Minimum Matched Fragment Ions is 6 (spectra containing less than 6 fragments of the same mass are not considered); Network TopK is set to 10; Maximum Connected Component Size is 100 (only when each cluster appears in 10 similar clusters each, the connection between clusters is maintained, i.e. the maximum range of a spectrum family is limited to 100 clusters).

[0065] SMART analysis: The NMR-HSQC data (HSQC spectrum) of C-1~C-3, E-1~E-3 and HD-C is collected on a Bruker Avance 600 DMX instrument, and is uploaded to the SMART platform after integration in a CSV (comma-separated value) file.

[0066] II. Results and Discussion 1. Construction and analysis of molecular network GNPS platform is an open access platform for processing mass spectrometry data, which integrates multiple mass spectrometry databases. Molecular network is one of the key functions of GNPS platform, which can realize the visualization of MS data. In the molecular network, each node contains specific mass spectrometry information. Based on the connection between the nodes of the molecules, the rapid identification of compounds can be realized. In the present application, the molecular network strategy was used to annotate the chemical components of Qinling honey. The mass spectrometry data of each extraction layer sample (including the methanol-soluble part of chloroform extraction layer HM-C, ethyl acetate extraction layer HM-E and n-butanol extraction layer HM-B) in positive and negative ion modes were imported into GNPS platform to complete the construction of molecular network.

[0067] The molecular network was compared and matched with the natural product database of GNPS platform, and the annotation results of the matched visualization nodes were obtained. The un-matched visualization nodes in the molecular network were annotated by using NAP annotation tool to predict their possible chemical skeleton and structure type.

[0068] The results showed that the molecular network formed in the positive ion mode contained 3310 nodes (including 20 nodes in the HM-C sample, 1090 nodes in the HM-E sample and 2200 nodes in the HM-B sample). Figure 2 After comparing with the MS / MS data in the GNPS spectral library, 20 nodes were annotated. Fatty acids and amides were the compound types with higher abundance in the identification results, including cis-9-Hexadecenoic acid, 15-HEDE, Linoleic acid, Linolenic acid, Petroselinic acid, 15-OxoEDE, 10-Hydroxydecanoic acid and Palmitamide, Octadecanamide, Methyl 2-benzamido-3-phenylpropanoate, N-Oleoylethanolamine, Linoleoyl ethanolamide, 13-Docosenamide, 9-Octadecenamide, N-Acetyl-L-phenylalanine, which were mainly derived from HM-C and HM-E. In addition, an alkaloid named dimethylisoalloxazine was identified, which is the main metabolite of vitamin B2 and is also considered to be the main reason for the bright yellow color of honey. The precursor ion m / z value of this node is 243.088, and the molecular formula of the compound is C 12 H10N4O2. Therefore, it is speculated that this node is the proton adduct of dimethylisoalloxazine ([M+H] +In negative mode, a network graph consisting of 1971 nodes was obtained. After comparison with MS / MS data in the GNPS spectral library, a total of 8 metabolites were identified. Figure 2 The B-group of the extract consists mostly of glycosides, primarily derived from HM-B, namely Sibiricose A3, Geniposidic acid, Kaempferol 7-neohesperidoside, Kaempferol-3-O-rhamnoside, and Loganic acid. Notably, ellagic acid, mainly derived from HM-E, was also identified, indicating that for Qinling honeybee honey, the extract obtained via n-butanol often contains glycosides, while chloroform and ethyl acetate are preferred for extracting aglycones. Information on all nodes identified by this method is shown in Table 2, with structures as follows: Figure 13 As shown.

[0069] Table 2. Overview of GNPS spectral library matching results As shown above, the chemical composition of Qinling honeybee honey includes typical components such as amides, fatty acids and their derivatives, polyphenols, and glycosides. The detected amide compounds include: Palmitamide, Octadecanamide, Methyl 2-benzamido-3-phenylpropanoate, N-Oleoylethanolamine, Linoleoyl ethanolamide, 13-Docosenamide, 9-Octadecenamide, and N-Acetyl-L-phenylalanine; fatty acids and their derivatives include: cis- 9-Hexadecenoic acid, 15-HEDE, Linoleic acid, Linolenic acid, Petroselinic acid, 15-OxoEDE, 10-Hydroxydecanoic acid, Monoolein, Octadecatrienoic acid methylester, Dimethyl sebacate, cis-7-Hexadecenoic acid methyl ester, Bis (2-ethylhexyl) adipate; Polyphenols are represented by Ellagic acid, Glycosides include: Sibiricose A3, Geniposidic acid, Kaempferol 7-neohesperidoside, Kaempferol-3-O-rhamnoside, Loganic acid. In addition, the matching results of the molecular network also contain a vitamin B2 major metabolite named dimethyl isoalloxazine.

[0070] 2. NAP annotation result analysis To further annotate the nodes without any library matching, different extraction layers were re-extracted and distilled, and the obtained fractions were analyzed by nuclear magnetic resonance. The SMART technology was applied to infer the specific structure type of the chemical components in the obtained fractions, which were different from the annotation results of the visual nodes and the possible chemical skeleton and structure type. The inference results were obtained. Specifically, the NAP annotation tool was used in the molecular network. Through the annotation of NAP, 25 additional nodes were determined as phenylpropanoids and polyketides in the positive ion mode, 45 nodes were identified as lipids and lipid molecules, 7 nodes were annotated as alkaloids, 19 nodes were annotated as organic acids, and the presence of benzene ring was detected in the structure of 43 nodes. It is worth mentioning that in the positive ion mode, based on the MS / MS data comparison in the GNPS library, the molecular cluster P-A (m / z 519.138) and P-B (m / z 371.101) were not matched to obtain the identified molecules. Figure 3 and Figure 5 ), in the annotation results of NAP, it was shown that it was a cluster of metabolites similar to the skeleton of phenylpropanoids or polyketides. In the structure annotation of the nodes with molecular weights of 519.138 and 371.101, respectively, the mother nucleus characteristics of coumarin were observed. At the same time, in the molecular cluster P-B (m / z 519.138) and P-A (m / z 371.101), Figure 3 and Figure 5 ), in the annotation results of NAP, it was shown that it was a cluster of metabolites similar to the skeleton of phenylpropanoids or polyketides. In the structure annotation of the nodes with molecular weights of 519.138 and 371.101, respectively, the mother nucleus characteristics of coumarin were observed. At the same time, in the molecular cluster P-B (m / z 519.138) and P-A (m / z 371.101),, GNPS library identified a node representing 9Z, 11E, 13E-octadecatrienoic acid methyl ester, and NAP annotation results classified this node as lipid and lipid molecule, and annotated the structures of three nodes with molecular weights of 285.242, 297.242 and 367.356 as lipids, respectively. Furthermore, nodes in molecular clusters P-C and P-D ( Figure 3 and Figure 5 ) were identified by NAP as lipids and lipid molecules, indicating that the nodes in this cluster might be lipids and lipid molecules or their derivatives. Finally, three nodes in molecular cluster P-E were annotated as alkaloids, which was similar to the matching results of GNPS, suggesting that molecular cluster P-E might be a cluster of alkaloids, and more likely to contain amide structures.

[0071] The molecular network in the negative ion mode was also annotated using NAP. Based on the annotation results of NAP, molecular cluster N-A ( Figure 4 and Figure 5 ) might be a cluster of phenylpropanoids and polyketides and their derivatives. Similar to the structural annotation in the positive mode, it was observed in the provided reference structures that they all contained the coumarin nucleus, except that more amide structures were contained here, while the structural annotation in the positive mode belonged to glycosides, which made up for the limitations of GNPS identification, because molecular cluster N-A was not annotated by GNPS library. In addition, the identification results of molecular clusters N-B and N-C ( Figure 4 and Figure 5 ) both showed the presence of alkaloids, with the difference being that the alkaloids in cluster N-B were more likely to contain amide structures, while the structures of cluster N-C were more similar to indole alkaloids. At the same time, the presence of lipids and lipid molecules was still annotated in the negative mode, and structural references were provided for some nodes in molecular clusters N-D and N-E ( Figure 4 and Figure 5 ), which were all fatty acids. Overall, in the negative mode, through the annotation of NAP, 13 nodes were annotated as phenylpropanoids and polyketides, 15 nodes were annotated as organic heterocyclic compounds, 8 nodes were annotated as lipids and lipid molecules, 2 nodes were annotated as organic acids, and in the structures of 4 nodes, the presence of benzene rings was found. Based on the above analysis, lipids and lipid molecules, amide compounds are the dominant chemical components of Qinling bee honey, while the identification results of phenolic acids and flavonoids and other compounds are not sufficient.

[0072] 3. SMART analysis results Based on SMART analysis, among the top 10 structures of HD-C cosine similarity scores, 8 were cyclic peroxides, and the highest score was a cembranoid diterpenoid compound named 7,8-Dihydroflabellatene B ( Figure 6), which indicates that there might be abundant oxides in HD-C and contain structures similar to macrocyclic diterpenes or cyclic oxides. According to the literature, 7,8-Dihydroflabellatene B has been found in organisms, and the appearance of similar structures in honey is speculated to be related to the honey swallow of bees. C-1~C-3 are sub-components obtained after silica gel column separation of HM-C, and the structure hypothesis generated by C-1 (0.08g) is mostly diterpenoids, including the highest scoring diterpenoid Rediocide C ( Figure 7 ). Similarly, the analysis results of C-2 (0.21g) also show diterpenoids, but unlike C-1, no macrocyclic lactone structure is found here ( Figure 8 ). In honey, terpenoids are usually detected as volatile ingredients, Mahmoud et al. (Mahmoud M A A, Kılıç B Ö, Aboul Fotouh M M, et al . Aroma active compounds of honey: analysis with GC-MS, GC-O, and molecular sensory techniques[J]. Journal of Food Composition and Analysis, 2024, 134: 106545.) comprehensively analyzed 187 different volatile compounds identified in 33 different types of honey, including terpenoids, aldehydes, alcohols, esters, acids, ketones and diketones, organic sulfides, lactones, phenols, furans, pyrazines and others, and pointed out that terpenoids and aldehydes are the most prominent categories. Kasiotis et al. (Kasiotis K M, Baira E, Iosifidou S, et alFingerprinting chemical markers in the mediterranean orange blossom honey: UHPLC-HRMS metabolomics study integrating melissopalynological analysis, GC-MS and HPLC-PDA-ESI / MS[J]. Molecules, 2023, 28(9): 3967.)People also said that in addition to hesperidin, Greek honey is richer in terpenes and iridoid compounds than flavonoids. Therefore, here, the present application believes that C-1 and C-2 may be clusters of terpenes, but C-1 is more likely to exist macrolide. Unlike C-1 and C-2, the structure of C-3 (0.18g) analyzed by SMART is all alkaloids ( Figure 9 ), and the parent nucleus feature of indole alkaloids appears. In fact, in the NAP annotation results about the negative mode, the node structure with a molecular weight of 186.056 derived from HM-C has been associated with indole alkaloids, and the analysis results of SMART further confirm the existence of chemical components related to indole alkaloid structure in Qinling bee honey. As one of the largest groups of plant secondary metabolites, alkaloids have been detected in various types of honey, including pyrrolizidine alkaloids, isoquinoline alkaloids, quinolone alkaloids, tropinane alkaloids, Gelsedine-type alkaloids, aconitine, oxymatrine, etc. (Wang Z, Zu T, Huang X, et al .Comprehensive investigation of the content and the origin of matrine-type alkaloids in Chinese honeys[J]. Food Chemistry, 2023, 402: 134254. ). Rarely, Protopine and Allocryptopine were used as potential markers to identify McRB honey (Truchado P, Martos I, Bortolotti L, et al. Use of quinoline alkaloids as markers of the floralorigin of chestnut honey[J]. Journal of Agricultural and Food Chemistry,2009, 57(13): 5680-5686. This inspired the inventors to believe that C-3 is a cluster of alkaloids, and it might be more closely related to indole alkaloids based on the results of SMART analysis.

[0073] HM-E was separated by semi-preparative separation to produce three sub-components, E-1 (0~11.5min, 0.18g) was eluted first on the reverse preparative column, and the analysis results showed that it might be mainly composed of glycosides (Table 2) Figure 10 ), especially 6α-Hydroxygeniposide with the highest cosine similarity score. This is consistent with the results of GNPS analysis in negative mode, in which Geniposidic acid was detected in HM-E based on GNPS library matching, and the structure of 6α-Hydroxygeniposide was very similar. In addition, GNPS annotated the node with m / z value of 431.098 derived from HM-E as Kaempferol-3-O-rhamnoside. This makes the inventors more believe that the main component of E-1 is glycosides represented by Geniposidic acid, which is similar to the structure of 6α-Hydroxygeniposide. In the analysis results of E-2 (11.5min~22min, 0.62g), there are 2 acridone alkaloids, 5 anthraquinones and 2 Phenylphenalenone compounds in the top 10 compounds with the highest cosine similarity score Figure 11), the cosine similarity score of Normelicopicine was the highest, and the anthraquinones were 1,3,5-Trihydroxy-2-(methoxymethyl)anthracene-9,10-dione, 1,3,6-tri-Hydroxy-2-methoxymethyl-9,10-anthraquinone, 6-Hydroxy-1,5-dimethoxy-2-(methoxymethyl) anthraquinone, Ophiohayatone A and 1,3-Dihydroxy-2-methoxymethylanthraquinone; the phenylphenalenones included 2,5,6-Trimethoxy-9-phenylphenalen-1-one and Haemoxiphidone, and the alkaloids all had acridone as the parent nucleus, which was not present in the C-3 hypothesis results. The GNPS analysis also seemed to support this conclusion, and in the GNPS results, the alkaloids were detected in both HM-E and HM-C, but the structures of Methyl 2-benzamido-3-phenylpropanoate mainly from HM-C and dimethylisoquinozine mainly from HM-E were not consistent, and dimethylisoquinozine was more suitable for clustering in E-2 than C-3. Therefore, the compounds present in E-2 were predicted to be anthraquinones and acridone alkaloids similar to anthraquinone structures. E-3 (0.76 g, 22 min~35 min) was the last sub-fraction eluted on the reverse preparation column, and long-chain hydrocarbons were present in the SMART-fitted compounds Figure 12 ), including long-chain fatty acids, esters and amides, and the highest scoring substance was alpha-alkyl butenolide dimer. It was speculated that the metabolites in E-3 might contain long-chain hydrocarbon structures, which was similar to the identification results of GNPS. In addition, the NAP annotation tool analyzed the nodes of m / z values of 293.247, 311.221 and 321.242 derived from HM-E as lipids and lipid molecules, and combined with the structure hypothesis of SMART, the chemical skeleton of these nodes might be similar to Paracaseolide A and Callyspongidic acid.

[0074] From the above, there are terpenoids (including macrolide structure), alkaloids (indole and acridone), long-chain lipids, glycosides and other structural types of compounds in Qinling Apis cerana cerana honey, represented by Rediocide C, Kopsiyunnanine C, Normelicopicine, Callyspongidic acid and 6α-Hydroxygeniposide respectively.

[0075] Finally, the annotation results of the visual nodes, possible chemical skeletons and structural types, and the inference results are scored for confidence, prioritized, and verified by retesting to finally determine the characteristic chemical component types of Qinling Apis cerana cerana honey. The five types of characteristic compounds, amide-fatty acid derivative-indole alkaloid-iridoid glycoside-anthraquinone, are formed by aligning the GNPS primary matching, NAP skeleton prediction and SMART fragment inference according to the retention time, and then quantifying the confidence and prioritizing. The nodes with scores less than 0.7 are upgraded by retesting the fractions, and only the qualified ones are locked.

[0076] To sum up, the chemical composition of Qinling Apis mellifera honey is analyzed by UPLC-Q-TOF-MS / MS, GNPS platform, NAP annotation tool and SMART technology. First, the Apis mellifera honey is extracted by chloroform, ethyl acetate and n-butanol in turn, and the mass spectrum information of each extraction layer is obtained, and then the mass spectrum information is integrated and visualized as a molecular network based on GNPS, and by comparing with the GNPS database, 28 metabolites are identified, including amides, fatty acids and their derivatives, polyphenols and glycosides, and it is indicated that the nodes constituting the same molecular cluster have similar chemical information, wherein the amides and fatty acids and their derivatives are the main detected components. Importantly, the detection results also include a kind of vitamin B2 main metabolite named dimethyl isoazoline, according to the research results of Du et al. (Du Y, Zhu H, Qiao J, et al. Characteristic components and authenticity evaluation of Chinese honeys from three different botanical sources[J]. Journal of Agricultural and Food Chemistry, 2023, 71(45): 17284-17294.), which may be the reason why the honey presents bright yellow. For the compounds that are not successfully identified, the structure is predicted by NAP annotation tool, and the results show that part of the nodes are identified as alkaloids, lipids and lipid molecules, phenylpropanoids and polyketides, organic heterocyclic compounds, which shows that there may be compounds related to these categories in the honey. These analysis results can show that alkaloids, lipids and lipid molecules may be the chemical components with higher abundance in Qinling Apis mellifera honey, and glycosides occupy a large advantage in content.

[0077] Based on the results of GNPS molecular network analysis, fractionation of HM-C and HM-E was performed by column chromatography and semi-preparative liquid chromatography, and the structure types of compounds in HD-C and each fraction were inferred based on the accurate identification of small molecules. HD-C was inferred to contain a large number of oxides, such as macrocyclic diterpenes or cyclic oxides. Among the fractions of HM-C, C-1 and C-2 were speculated to be clusters of terpenoids, but C-1 was more likely to contain a macrocyclic lactone structure similar to Rediocide C, and the structure type of the compounds in C-2 was more similar to Jiadifenoic acid. The main component of C-3 was identified as an alkaloid, and was most likely similar to indole alkaloid structures, represented by Kopsiyunnanine C. Among the sub-components of HM-E, the cosine similarity score of Normelicopicine was the highest in the analysis results of E-2, and there were 5 anthraquinones in the top 10 structures, suggesting that E-2 contains anthraquinones and acridinones, which are different from the alkaloid structures in C-3. This finding supplements the annotation results of NAP annotation tool for alkaloids in HM-C and HM-E. The analysis results of E-3 show that there may be long-chain aliphatic hydrocarbon structures, including long-chain fatty acids, esters and amides, which not only conform to the matching results of GNPS, but also infer that the annotation of lipids and lipid-like molecules in HM-E by NAP tool is similar to Paracaseolide A and Callyspongidic acid structures. Finally, E-1 was inferred to be a cluster of glycosides represented by Geniposidic acid, which is similar to the 6α-Hydroxygeniposide structure, similar to the matching results of GNPS in the negative mode.

[0078] The analysis strategy combining GNPS and NAP annotation tool reduces the dependence on databases to some extent compared to traditional analysis methods. The analysis results of SMART technology are not only similar to the GNPS database matching and NAP annotation results, but also provide more specific explanations and descriptions of some NAP annotation results. The discovery of terpenoids (including macrocyclic lactone structures) provides key clues for the analysis of uncharacterized compound structures. In summary, the analysis based on SMART not only overcomes the limitations of mass spectrometry-based analysis due to sample solubility, but also makes the prediction of the structure type of compounds in total extract or fractions more concrete. This analysis method is rapid and lays a theoretical foundation for quickly identifying the structure type of metabolites in honey crude extract and fractions.

[0079] It should be noted that when numerical ranges are involved in the present application, it is understood that each intervening value, to the left and right of the recited ranges, is encompassed within the present application. For example, if a range is stated as 1 to 50, it is intended that any number between 1 and 50 is also encompassed within this application.

[0080] While the preferred embodiments of the application have been described, additional variations and modifications can be made to these embodiments by those skilled in the art once they have the benefit of the present disclosure without departing from the spirit and scope of the application.

Claims

1. A method for analyzing the characteristic chemical components of honey from the central peak of the Qinling Mountains, characterized in that, Includes the following steps: Qinling honeybee honey was extracted using different organic solvents to obtain different extraction layers, and mass spectrometry data of the chemical components in the different extraction layers were obtained. Based on the GNPS platform, the mass spectrometry data of chemical components in different extraction layers are clustered and integrated, and visualized as a molecular network. The molecular network is compared and matched with the natural product database of the GNPS platform to obtain the annotation results of the successfully matched visualization nodes; the unmatched visualization nodes in the molecular network are annotated using the NAP annotation tool to predict their possible chemical skeleton and structure type. Different extraction layers were subjected to further extraction and distillation. The resulting fractions were analyzed by NMR. Using SMART technology, the specific structural types of chemical components in the fractions that differed from the annotation results of the visualization nodes and from the possible chemical skeletons and structural types were inferred, yielding the following results: The annotation results of the visualized nodes, the possible chemical skeletons and structural types, and the inference results are evaluated for confidence scores, priority decisions, and retesting to ultimately determine the characteristic chemical component types of Qinling Zhongfeng honey.

2. The analytical method according to claim 1, characterized in that, The characteristic chemical components of the Qinling Zhongfeng honey include amides, fatty acids and their derivatives, glycosides, terpenes and anthraquinones.

3. The analytical method according to claim 1, characterized in that, The steps to obtain the molecular network are as follows: The mass spectrometry data of chemical components in the different extraction layers are converted into .mzXML format data files, and the converted .mzXML format data files are uploaded to the GNPS platform. The GNPS platform automatically executes the process of "spectral redundancy removal - molecular fingerprint extraction - full-to-full similarity calculation - force-directed network layout" to complete the construction of the molecular network.

4. The analytical method according to claim 3, characterized in that, The converted .mzXML data file is imported into the Feature-Based Molecular Networking workflow of the GNPS platform. The mass deviation between the parent ion and fragment ions is set to 0.02 Da, the similarity threshold is ≥0.7, the minimum number of matching fragment ions is 6, the minimum cluster size is 2, the number of TopK edges is set to 10, and the maximum number of connected components is set to 100. The GNPS platform automatically completes redundancy removal, similarity calculation, TopK pruning, and giant cluster segmentation, and finally outputs an interactive molecular network map.

5. The analytical method according to claim 1, characterized in that, The different organic solvents include chloroform, ethyl acetate and n-butanol, and the corresponding different extraction layers include chloroform extraction layer, ethyl acetate extraction layer and n-butanol extraction layer.

6. The analytical method according to claim 5, characterized in that, Before extracting Qinling honey with chloroform, ethyl acetate and n-butanol in sequence, the Qinling honey was dispersed in an aqueous solution containing 2% NaCl by mass volume. The mass ratio of the Qinling honey to the volume ratio of the NaCl aqueous solution is 2.25 kg: 1 L~3 L.

7. The analytical method according to claim 5, characterized in that, The different extraction layers were subjected to further extraction and distillation, and the resulting fractions were analyzed by nuclear magnetic resonance. Simultaneously, the SMART technique was applied to infer the specific structural types of unidentified compounds in the resulting fractions. The steps to obtain the inference results are as follows: The chloroform extract layer is dissolved in methanol to obtain a methanol-insoluble fraction and a methanol-soluble fraction of the chloroform extract layer; the methanol-soluble fraction of the chloroform extract layer is separated by chromatography to obtain multiple fractions corresponding to the methanol-soluble fraction of the chloroform extract layer. The ethyl acetate extract layer was separated by chromatographic method to obtain multiple fractions corresponding to the ethyl acetate extract layer; The methanol-insoluble fraction of the chloroform extraction layer, the corresponding fractions of the methanol-soluble fractions of the plurality of chloroform extraction layers, and the corresponding fractions of the plurality of ethyl acetate extraction layers were dissolved in a deuterated solvent and the NMR data were collected. The SMART-NMR platform was used to automatically extract peaks, cluster chemical shifts, and identify skeleton fragments from the NMR data of each of the above fractions, generating a prediction list of the assumed chemical skeleton and / or structure type corresponding to each fraction. The prediction list is associated with unannotated or low-matching visualization nodes in the molecular network: if the exact mass deviation between the parent ion m / z of a visualization node and the assumed skeleton in the prediction list is ≤5ppm, and the MS / MS fragments of the visualization node match the main theoretical fragments of the assumed chemical skeleton ≥4, then the visualization node is annotated as the corresponding skeleton or structure type, thereby inferring the specific structure type of the unidentified compounds in the obtained fraction.

8. The analytical method according to claim 1, characterized in that, Mass spectrometry data of chemical components in different extraction layers were obtained by detecting the different extraction layers using ultra-high performance liquid chromatography-quadrupole-time-of-flight tandem mass spectrometry. The chromatographic conditions for detection are as follows: The Thermo Scientific™ Q Exactive Plus quadrupole-electrostatic track trap high-resolution mass spectrometry system was used; the chromatographic column was an ACQUITY UPLC BEH C18 column with dimensions of 2.1 × 50 mm and a diameter of 1.7 µm; the column temperature was 30 °C; the flow rate was 0.20 mL / min; and the injection volume was 3.0 μL. The mobile phase consists of mobile phase A and mobile phase B; mobile phase A is 0.1% formic acid water by volume, and mobile phase B is acetonitrile.

9. The analytical method according to claim 8, characterized in that, The chromatographic conditions for detection also include the elution procedures shown in the table below:

10. The analytical method according to claim 8, characterized in that, The mass spectrometry conditions for detection are as follows: The experiment was conducted in both positive and negative ion modes, equipped with an electrospray ionization source, with an atomizer pressure of 35 psi; sheath gas temperature of 350 °C and flow rate of 10 AU; auxiliary gas flow rate of 1.0 AU; capillary temperature of 320 °C; and MS scan range of 100. m / z ~1500 m / z MS / MS range is 50 m / z ~1500 m / z The collision energy is 20eV, 40eV or 60eV.

Images