Method for identifying nutmeg crude product and bran stewed product and application thereof

Abstract

The invention belongs to the technical field of traditional Chinese medicine identification, and particularly relates to a method for identifying a nutmeg crude product and a bran stewed product and application of the method. According to the method for identifying the nutmeg crude product and the bran stewed product, provided by the invention, 5-hydroxymaltol, maltol, N-methyltryptamine, adipic acid and myristicin B1 are taken as symbolic components, and a to-be-detected sample is detected by using ultra-high performance liquid chromatography-ultraviolet detection / quadrupole-time-of-flight mass spectrometry; and judging whether the to-be-detected sample is a nutmeg crude product or a bran stewed product according to the content of the symbolic component. The symbolic component obviously changes before and after the myristica fragrans is processed, the myristica fragrans bran-stewed product processed in a laboratory or sold in the market can be identified by detecting the content of the symbolic component, and a practical method is provided for distinguishing the myristica fragrans raw product from the bran-stewed product.

Description

Technical Field

[0001] This invention belongs to the field of traditional Chinese medicine identification technology, specifically relating to a method for identifying raw nutmeg and bran-processed nutmeg, and its application. Background Technology

[0002] nutmeg( Myristica fragrans Nutmeg (known as nutmeg in Traditional Chinese Medicine) is a traditional spice widely used in food, beverages, and herbal preparations. Its chemical composition includes volatile oils, fixed oils, lignans, flavonoids, triterpenes, starch, proteins, and trace oligosaccharides, exhibiting multiple pharmacological activities such as regulating the digestive system, lowering blood sugar, analgesia, anti-inflammation, anti-oxidation, and anti-cancer effects. However, raw nutmeg contains toxic components such as myristole, which may cause hepatotoxicity. Therefore, processing techniques such as simmering in wheat bran are necessary to reduce toxicity and enhance its antidiarrheal and anti-inflammatory effects. Although simmered nutmeg has significant value in cooking and personal care, current research has found a high degree of similarity in fingerprint profiles between raw and processed products, necessitating the use of advanced metabolomics techniques to reveal subtle chemical differences. Summary of the Invention

[0003] To address the above technical problems, this invention provides a method for distinguishing between raw nutmeg and bran-processed nutmeg, and its application. This method uses five non-volatile components as markers, combined with ultra-high performance liquid chromatography-ultraviolet detection / quadrupole-time-of-flight mass spectrometry to identify the sample, enabling accurate differentiation between raw and bran-processed nutmeg.

[0004] To achieve the above-mentioned objectives, the present invention adopts the following technical solution: In a first aspect, the present invention provides a method for distinguishing between raw nutmeg and bran-processed nutmeg. The method uses 5-hydroxymaltol (CAS Registry No.: 1073-96-7), maltol (CAS Registry No.: 118-71-8), dipterine (CAS Registry No.: 61-49-4), adipic acid (CAS Registry No.: 124-04-9), and fragransin B1 (CAS Registry No.: 112516-03-7) as marker components. The sample is then analyzed using ultra-high performance liquid chromatography-UV detection / quadrupole-time-of-flight mass spectrometry (UHPLC-UV / QTOF-MS). The content of these marker components determines whether the sample is raw or bran-processed nutmeg.

[0005] The present application is based on the research of non-targeted metabolomics combined with machine learning model, and 5 marker components with significant changes before and after the processing of mace are mined, which can be used to identify mace crude products and roasted products: 5-hydroxymaltol, maltol, adipic acid and myristicin B1 increase after processing, while the content of N-methyl tryptamine decreases. Based on the content detection of the above 5 marker components, mace crude products and roasted products can be distinguished.

[0006] Preferably, the sample to be tested is the water extract of mace crude product or the water extract of roasted product.

[0007] Further preferably, the method for preparing the water extract of mace crude product or the water extract of roasted product is: after water reflux extraction of mace crude product or roasted product, solid-liquid separation is performed, and the water extract is taken.

[0008] More preferably, the water reflux extraction time is 1 hour, and the soaking time before water reflux extraction is 30 minutes; the solid-liquid separation method is 20,817 x g centrifugation for 10 minutes, and the supernatant is taken.

[0009] Preferably, the chromatographic conditions of ultra-high performance liquid chromatography-ultraviolet detection / quadrupole-time-of-flight mass spectrometry are as follows: Chromatographic column: octadecylsilane bonded silica gel chromatographic column; Mobile phase A: 0.05%-0.15% v / v formic acid aqueous solution, and mobile phase B: acetonitrile, linear gradient elution is performed, and the program of the linear gradient elution is as follows: 0-22 minutes, 5%-42% B; 22-26 minutes, 42% B; 26-30 minutes, 42%-55% B; 30-37 minutes, 55%-80% B; 37-43 minutes, 80%-95% B; Flow rate: 0.25-0.35 mL / min; Column temperature: 35-45 ℃.

[0010] The above mobile phase, column temperature, flow rate and elution program used in the present application can obtain ideal separation effect, so as to realize accurate detection of the above five marker components.

[0011] Further preferably, the chromatographic column is selected from: CORTECS C18, specification: 2.1 x 100 mm, 1.6 μm; or CORTECS UPLC C18+, specification: 2.1 x 100 mm, 1.6 μm; or HSS T3, specification: 2.1 x 100 mm, 1.8 μm; or BEH C8, with dimensions of 2.1 × 100 mm and a micrometer diameter of 1.7 μm; or BEH C18, with dimensions of 2.1 × 100 mm and a micrometer diameter of 1.7 μm; or CSH C18, with dimensions of 2.1 × 100 mm and a micrometer diameter of 1.7 μm; or CSH Phenyl-Hexyl, with specifications of 2.1 × 100 mm and 1.7 μm; or CSH Cyano, with dimensions of 2.1 × 100 mm and 1.8 μm; or CSH Fluoro-Phenyl, with specifications of 2.1 × 100 mm and 1.7 μm; or Atlantis Premier BEH C18 AX, dimensions 2.1 × 100 mm, 1.7 μm; or Zorbax SB-Aq, with dimensions of 2.1 × 100 mm and a micrometer diameter of 1.8 μm; or Zorbax Extend C18, with dimensions of 2.1 × 100 mm and a thickness of 1.8 μm; or Kinetex 1.7μ Biphenyl, with a specification of 2.1×100 mm and 1.7 μm; or Kinetex XB-C18, with dimensions of 2.1 × 100 mm and a micrometer diameter of 1.7 μm; or Luna Omega 1.6μ Polar C18, with dimensions of 2.1×100 mm and 1.6 μm.

[0012] More preferably, the chromatographic column is a Luna Omega Polar C18 with dimensions of 2.1 × 100 mm and a diameter of 1.6 μm. Under the chromatographic conditions of this invention, this column can yield the highest number of resolvable peaks.

[0013] More preferably, the mobile phase A is 0.1%. v / v Formic acid aqueous solution.

[0014] More preferably, the flow rate is 0.3 mL / min.

[0015] More preferably, the column temperature is 40 °C.

[0016] Preferably, the ultra-high performance liquid chromatography-ultraviolet detection / quadrupole-time-of-flight mass spectrometry uses positive ion mode ESI for mass spectrometry acquisition, and the scanning range is [not specified]. m / z100-800; Mass spectrometry conditions included: capillary voltage 4000 V, nozzle voltage 500 V, fragmentor voltage 380 V, drying gas temperature 200 °C, drying gas flow rate 12 L / min, nebulizer pressure 35 psi, sheath gas temperature 350 °C, and collision energy set to 5–30 V.

[0017] Secondly, the present invention also provides the application of the above method in constructing a machine learning model for identifying raw nutmeg and its bran-based products.

[0018] By using the above identification methods to detect the five marker components, and combining them with a machine learning model, it is possible to effectively distinguish between raw nutmeg and its bran-processed products.

[0019] Preferably, the machine learning algorithm of the machine learning model includes support vector machine (SVM), partial least squares discriminant analysis (PLS-DA), and random forest (RF).

[0020] More preferably, the machine learning algorithm is a support vector machine or a random forest. Support vector machines and random forests have high accuracy in identifying nutmeg bran products.

[0021] More preferably, the machine learning algorithm is a support vector machine. Experiments have shown that by using the above analytical conditions to detect the five marker components and then using the SVM model, 90% of commercially available nutmeg bran products can be identified.

[0022] The beneficial effects of this invention are as follows: Through experimental research and screening, this invention has discovered five important marker components related to the nutmeg roasting process: 5-hydroxymaltol, maltol, N-methyltryptamine, adipic acid, and myristicin B1. These marker components help identify laboratory-processed or commercially available nutmeg bran-roasted products, providing a practical method for distinguishing between raw nutmeg and bran-roasted products. Attached Figure Description

[0023] Figure 1The following are the UHPLC / QTOF-MS total ion chromatograms of representative raw nutmeg samples (RN17), homemade braised nutmeg samples (BN-L6), and commercially available braised nutmeg samples (BN-M17) from Example 1; peaks S1, S2, S5, S9, S10, S13, S14, and S18 represent catechin, vanillin, 4-allyl-2,6-dimetoxyphenol, fragransin B1, odoratisol A, myristicin, myrislignan, and malabaricone B, respectively; peaks M1, M2, M3, and M4... M5 and M5 are 5-hydroxymaltol, maltol, dipterine, adipic acid, and fragransin B1, respectively. Figure 2 The following are the chemometric analysis results of non-volatile components in the raw and homemade braised nutmeg samples from Example 1. A: Principal component analysis score plot; B: Partial least squares discriminant analysis score plot; C: Volcano plot including VIP values; Figure 3 The results of screening for processing-related biomarkers of non-volatile components of nutmeg using three machine learning models in Example 1 are shown. A: ROC curve, average AUC value, and 95% confidence interval (CI); B: Confusion matrix of external validation of the three models; C: Feature importance ranking of the top 15 important features in the SVM and RF models; D: Venn diagram of the top 7 feature variables in the SVM and RF models. Figure 4 This section shows the validation results of five non-volatile markers using an external validation dataset in Example 1. A: ROC curve, mean AUC values ​​and 95% confidence intervals of SVM and RF validation tests; B: Confusion matrix of SVM and RF external validation; C: Heatmap; D: Box plot showing content differences; Figure 5 This is the result of the identification of 20 batches of commercially available bran-baked nutmeg samples based on five markers in Example 1. Red marks indicate identification errors; Figure 6 This refers to the fragmentation mode of the five markers in Example 1; Figure 7 The total ion flow charts obtained using different flow patterns in Comparative Example 1 are shown below. Figure 8The total ion chromatograms for the reversed-phase ultra-high performance liquid chromatography columns used in Comparative Example 2 were obtained using Atlantis Premier BEH C18 AX, Luna Omega 1.6μ Polar C18, CSH C18, CORTECS UPLC C18+, and Kinetex 1.7μ Biphenyl columns. Figure 9 The total ion chromatograms for the reversed-phase ultra-high performance liquid chromatography columns used in Comparative Example 2 were obtained using CSH Phenyl-Hexyl, Zorbax SB-Aq, HSS T3, BEH C8, and Kinetex XB-C18 columns. Figure 10 The total ion chromatograms for the reversed-phase ultra-high performance liquid chromatography columns used in Comparative Example 2 were obtained using Zorbax Extend C18, CORTECS C18, CSH Cyano, BEH C18, and CSH Fluoro-Phenyl. Figure 11 The total ion chromatograms for Comparative Example 2 were obtained using reversed-phase ultra-high performance liquid chromatography columns such as Zorbax Eclipse Plus C18, BEHShield RP18, HSS C18 SB, Kinetex 1.7μ EVO C18, and CORTECS T3. Detailed Implementation

[0024] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. All other embodiments obtained by those skilled in the art based on the implementation methods of this invention without inventive effort fall within the protection scope of this invention.

[0025] Nutmeg requires processing techniques such as roasting with wheat bran to reduce its toxicity and enhance its antidiarrheal and anti-inflammatory effects. Fingerprint spectroscopy is a commonly used tool for identifying traditional Chinese medicine, but the fingerprint spectra of raw and processed nutmeg are highly similar, making accurate identification of raw and roasted nutmeg difficult. Furthermore, this invention has found that while traditional chemometric methods can identify many differential metabolites, their performance in classification prediction is not ideal. In addition, this invention also attempted to establish an analytical method based on volatile components, but the results showed that its classification performance was relatively poor.

[0026] To address the above issues, this invention utilizes a combination of non-targeted metabolomics and machine learning models to identify five marker components that can be used to distinguish between raw and bran-processed nutmeg. These marker components exhibit significant changes before and after nutmeg processing. Based on these findings, this invention provides a method for distinguishing between raw and bran-processed nutmeg. Using 5-hydroxymaltol, maltol, N-methyltryptamine, adipic acid, and myristicin B1 as marker components, the method employs ultra-high performance liquid chromatography-ultraviolet detection / quadrupole-time-of-flight mass spectrometry to detect the sample. The content of these marker components determines whether the sample is raw or bran-processed nutmeg.

[0027] This invention also provides the application of the above method in constructing a machine learning model for identifying raw nutmeg and its bran-processed products.

[0028] The following detailed description, using specific embodiments, will further illustrate this point.

[0029] The reagents and chemicals used in the following examples: The origin information for 60 batches of nutmeg samples (raw RN, homemade braised nutmeg BN-L, and commercially available braised nutmeg BN-M) is detailed in Table 1. Acetonitrile, methanol (Fisher, USA), and formic acid (FA; Sigma-Aldrich, USA) used for chromatographic analysis were all LC-MS grade. Ultrapure water was prepared using a Milli-Q Integral 5 system (Millipore, USA).

[0030] Table 1. Detailed information on 60 batches of nutmeg samples

[0031] Preparation method of homemade bran-baked products: Take raw nutmeg (RN) and bran together and bake: heat at 150-160 °C for 15 minutes until the bran is brownish-yellow and the nutmeg surface is brownish-red and cracks appear. Sift out the bran while hot, and immediately grind through a 40-mesh sieve after cooling.

[0032] Unless otherwise specified, all other reagents and medicinal materials used in the following examples are commercially available or obtained using methods known in the art. Unless otherwise specified, all experimental methods used in the following examples are conventional methods known in the art.

[0033] Example 1 This invention provides a method for distinguishing between raw nutmeg and bran-processed nutmeg, and its application in constructing a machine learning model for distinguishing between raw nutmeg and bran-processed nutmeg.

[0034] 1. Preparation of the sample to be tested Accurately weigh 5 g of the raw or bran-processed nutmeg sample to be tested, add 50 mL of ultrapure water, soak for 30 minutes, and then reflux for 1 hour. The extract was analyzed at 20,817 × 10⁻⁶. g Centrifuge for 10 minutes, and take the supernatant as the sample to be tested for UHPLC / QTOF-MS analysis.

[0035] Each nutmeg sample in Table 1 was extracted and centrifuged according to the above-mentioned sample preparation method. Then, all the samples were mixed in equal volumes as quality control samples (QC) and analyzed by UHPLC / QTOF-MS.

[0036] 2. UHPLC / QTOF-MS detection The detection was performed using an Agilent 1290 Infinity II ultra-high performance liquid chromatography system connected in series with an Agilent 6550 quadrupole time-of-flight mass spectrometer (Agilent Technologies, USA).

[0037] Chromatographic conditions: The chromatographic column was a Luna Omega Polar C18 (2.1 × 100 mm, 1.6 μm; Phenomenex, USA); the column temperature was 40 ℃; and the mobile phase was 0.1%. v / v Formic acid aqueous solution (A) - acetonitrile (B) at a flow rate of 0.3 mL / min, with the following gradient elution program: 0–22 min, 5%–42% B; 22–26 min, 42% B; 26–30 min, 42%–55% B; 30–37 min, 55%–80% B; 37–43 min, 80%–95% B. The equilibration time was 3 min, and the injection volume was 2 μL.

[0038] Mass spectrometry acquisition uses positive ion mode ESI, scan range m / z 100-800. Key parameters were set as follows: capillary voltage 4000 V, nozzle voltage 500 V, fragmentor voltage 380 V, drying gas temperature 200 °C, drying gas flow rate 12 L / min, nebulizer pressure 35 psi, and sheath gas temperature 350 °C. To elucidate the structural characteristics of the identified differential compounds, targeted MS / MS analysis was performed to obtain their secondary mass spectra, with collision energies set to 5–30 V.

[0039] Table 1 shows the UHPLC / QTOF-MS total ion chromatograms of representative raw nutmeg sample (RN17), homemade braised nutmeg sample (BN-L6), and commercially available braised nutmeg sample (BN-M17) from the 60 batches of nutmeg samples. Figure 1 As shown.

[0040] 3. Comparative analysis and identification of non-volatile components showing significant differences between raw RN and homemade bran-based product BN-L. Raw UHPLC / QTOF-MS data were preprocessed using Agilent Profinder software, including noise filtering, baseline correction, deconvolution, peak alignment, peak identification, and feature extraction. To eliminate solvent peak interference, data with retention times of 0-3 min were removed; metabolites with detection rates below 80% in each group were also removed to improve data quality. Subsequent analysis was performed using MetaboAnalyst commercial software (https: / / www.metaboanalyst.ca): missing values ​​were filled using the KNN algorithm (k-nearest neighbor method), data were transformed using square root, and Pareto normalization was performed.

[0041] Differential metabolite screening employed univariate analysis (T-test and fold change analysis) combined with chemometric methods (principal component analysis (PCA) and orthogonal partial least squares discriminant analysis (OPLS-DA)).

[0042] The constructed unsupervised PCA model helps assess the dispersion of metabolomic data for nutmeg and its processed products (RN and BN-L), such as... Figure 2 As shown in Figure A, the contribution rate of the first principal component (PC1) is 14.3%, and the contribution rate of the second principal component (PC2) is 10.3%, which reflects 24.6% of the information between the two samples. This indicates a high degree of metabolomics similarity among all tested nutmeg samples. Quality control data (see...) Figure 2 The QC clustering in Figure A is good. Analysis of the raw data revealed that the relative standard deviation (RSD) variation of over 89.07% of the data points was ≤ 30%. This demonstrates the instrument's stability and allows for further data analysis.

[0043] The score graph of OPLS-DA is as follows Figure 2 As shown in Figure B, the RN and BN-L samples are well distinguishable. The prediction parameters of the OPLS-DA model are R. 2 Y 0.989, R 2 X 0.257, Q 2 A p-value of 0.611 and a p-value <0.05 indicate that the model is stable and reliable. The variable importance projection (VIP) score obtained by the OPLS-DA method can be used as an indicator to assess the importance of metabolites in distinguishing different groups. The higher the VIP value, the more significant the role of the metabolite in model construction and the stronger its ability to distinguish between different groups.

[0044] Chemometric analysis identified 57 differentially expressed compounds with VIP values ​​≥ 1, FC > 2 or < 0.5, and P < 0.05. A volcano plot of metabolites containing VIP values ​​(e.g.) Figure 2 Figure C in the table shows that 29 compounds were significantly upregulated and 28 compounds were significantly downregulated. High-resolution MS analysis yielded... 2 The data allowed for the identification or preliminary identification of 18 compounds, as shown in Table 2.

[0045] Table 2. List of 18 identified non-volatile differential metabolites between the RN and BN-L groups.

[0046] Note: * indicates identification by comparison with a reference standard.

[0047] Feature selection and model performance evaluation were performed on the MetaboAnalyst platform. Three supervised machine learning algorithms (Support Vector Machine (SVM), Partial Least Squares Discriminant Analysis (PLS-DA), and Random Forest (RF)) were employed. ROC curves were constructed based on data from 57 differentially expressed metabolites, and model validation was performed using Monte Carlo Cross-Validation (MCCV). In each iteration, two-thirds of the samples were randomly selected as the training set to evaluate feature importance, and the remaining one-third was used as an independent validation set to test model performance. This process was repeated to obtain stable estimates of classification accuracy, and the 95% confidence intervals for the AUC values ​​of each model were calculated. The results are as follows: Figure 3 As shown in Figure A, on the validation set, the average AUC values ​​of PLS-DA, SVM, and RF all exceeded 0.85, with SVM and RF achieving near-perfect accuracy (approximately 100%). This indicates that these three machine learning models all demonstrated excellent discriminative ability in distinguishing between RN and BN-L.

[0048] To further compare the evaluation capabilities of these three machine learning models, 20 batches of BN-M samples were used as the test set, and external validation was performed based on 57 differentially expressed metabolites (i.e., validation using commercially available bran-roasted nutmeg). For discrimination performance on the test set, the accuracy of all three models was greater than 0.65, significantly lower than the 0.85 achieved by other models. This lower accuracy can be attributed to the influence of processing conditions employed by different manufacturers. In contrast, SVM and RF demonstrated better classification performance (75% accuracy), while PLS-DA was the worst (65% accuracy). Figure 3 (As shown in Figure B). Further, SVM and RF are used to filter for process-related markers. (See Figure B in the original text.) Figure 3As shown in Figure C, the first seven components of SVM and the first six components of RF achieve higher discriminative power, reaching prediction accuracies of 85% and 75%, respectively. Intersection analysis of the selected first seven variables using Venn diagrams was performed to offset the bias of specific algorithms, yielding five common markers (such as...). Figure 3 (As shown in Figure D) Using the content (integral peak area) of five markers, classification models were developed using two selected machine learning methods to evaluate their discrimination accuracy. These models demonstrated that: (1) the accuracy in identifying BN-L and RN exceeded 90% in multiple iterations on the validation set; and (2) the average ROC curve area was close to 1 (e.g., Figure 4 As shown in Figure A). External validation further confirmed the reliability of the model, with SVM (90% accuracy) outperforming RF (85% accuracy) (as shown in Figure A). Figure 4 (As shown in Figure B). The difference in content between RN and BN-L samples is shown in Figure B. Figure 4 As shown in Figures C and D, this confirms that the five markers can distinguish the processing state of nutmeg. The content of Markers 1, 2, 4, and 5 increases after processing, while the content of Marker 3 decreases. Figure 5 The final prediction results of an SVM using five markers for 20 batches of BN-M samples are shown.

[0049] 4. Structural identification of five types of markers The structure of five markers was identified using MetaboAnalyst software.

[0050] Marker 1 ( t R 3.12 min, C6H6O4) generated through collision-induced dissociation (CID) m / z [M+H] of 143.0337 + Precursor ions and m / z 125.0232 ([M+H-H2O]) + ) and 69.0334 ([M+H-H2O-2CO] + The corresponding fragment ions of ) were identified as 5-hydroxymaltol.

[0051] Marker 2 ( t R 3.32 min, C6H6O3) produced m / z The precursor ion of 127.0386 can generate m / z 109.0284 ([M+H-H2O]) +) and 81.0334 ([M+H-H2O-CO] + The product ion of ) was identified as maltol by comparison with a reference standard.

[0052] According to the PubMed public database, barley ( Hordeum vulgare L.) and common wheat ( Triticum aestivum L.) is associated with both of these compounds. Bran is the hard outer shell that detaches during flour milling. These two substances were detected in the aqueous decoction of treated nutmeg samples because bran residue adhered to the outer skin of the wheat kernels during processing.

[0053] When analyzing secondary fragment ions, Marker 3 was most likely N-methyltryptamine (dipterine), a compound previously reported to be associated with nutmeg. This compound contains an indole core and belongs to the tryptophan alkaloid family. It is readily decomposed by heat, hence its higher concentration in unprocessed nutmeg compared to processed products.

[0054] Marker 4 ( t R At 3.19 min, C4H7N3O2) produced a protonated dehydration precursor ion with a mass-to-charge ratio of 129.0542. m / z Fragments with values ​​of 111.0440, 101.0597, and 83.0491 were detected and classified as [M+H-2H2O]. + [M-H2O+H-CO] + and [M+H-2H2O-CO] + The platform predicts that Marker 4 is most likely adipic acid, but no source information is available. Its fragmentation information is consistent with information retrieved from online databases (Human Metabolomics Database, HMDB; and MassBank Europe).

[0055] Marker 5 ( t R 18.15 min, C 22 H 28 O7) shows [M-H2O+H] + The precursor ion, with a mass-to-charge ratio of 387.1794, was analyzed by CID-MS / MS to produce... m / z Fragment ions of 153.0546 and 219.1057 are due to bond breaking of 2,6-dimethoxyphenol. m / z Fragment ions at 219.1057 generate a secondary ion due to the loss of acetylene. m / z193.0859. m / z The fragment ion 179.0702 is produced by the cleavage of 3,4-dimethyltetrahydrofuran. Removal of 2,5-dimethoxycyclopentadien-1-ol from the precursor ion yields a... m / z Fragment ions at 233.1172. Platform matching Marker 5 is fragransin B1, a lignan compound derived from nutmeg.

[0056] The fragmentation mode of these five markers is as follows: Figure 6 As shown.

[0057] Comparative Example 1 This comparative example provides the detection results using different mobile phases in UHPLC / QTOF-MS detection.

[0058] Sample to be tested: Quality control sample from Example 1.

[0059] Chromatographic conditions: Chromatographic column: Waters ACQUITY UPLC HSS T3 column (2.1 × 100 mm, 1.8 µm); column temperature: 40℃; mobile phase: 0.1% v / v Formic acid solution (A) - methanol (B), 0.1% v / v Formic acid solution (A) - acetonitrile (B), water (A) - acetonitrile (B), water (A) - 0.1% v / v Acetonitrile formate (B), 0.1% v / v Formic acid solution (A) - 0.1% v / v Formic acid acetonitrile (B), flow rate 0.3 mL / min, gradient elution program as follows: 0-12 min, 15%-42% B; 12-18 min, 42% B; 18-22 min, 42%-55% B; 22-30 min, 55%-80% B; 30-32 min, 80%-95% B. Injection interval equilibration time 3 min, injection volume 1 μL.

[0060] Sample separation effect as follows Figure 7 As shown. When acetonitrile is used as the organic phase, water-acetonitrile or 0.1% is employed. v / v Using formic acid-acetonitrile as the mobile phase can resolve more peaks in the total ion chromatogram. A 0.1% solution was used. v / v Using formic acid-acetonitrile as a mobile phase can result in a smoother baseline. (0.1%...) v / v When formic acid is in aqueous phase, using acetonitrile as the organic phase can yield more peaks.

[0061] Comparative Example 2 This comparative example provides the detection results using different chromatographic columns in UHPLC / QTOF-MS detection.

[0062] Sample to be tested: Quality control sample from Example 1.

[0063] Chromatographic conditions: Chromatographic column: as shown in Table 2; column temperature: 40 ℃; mobile phase: 0.1% v / v Formic acid water (A) - acetonitrile (B) at a flow rate of 0.3 mL / min, using a gradient elution program (as shown in Table 3) to achieve the best separation effect for each column, with an injection interval equilibration time of 3 min and an injection volume of 1 μL.

[0064] Table 2 Column Information

[0065] Table 3 Gradient elution program

[0066] Sample separation effect as follows Figure 8~Figure 11 As shown, BEH Shield RP18 and Zorbax Eclipse Plus C18 exhibit large baseline variations in the total ion chromatogram, while Kinetex EVO C18, HSS C18 SB, and CORTECS T3 show poor peak shapes. The remaining columns can achieve good separation of nutmeg components, with the Luna Omega 1.6μ Polar C18 column providing the most resolvable peaks.

[0067] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for distinguishing between raw nutmeg and bran-baked nutmeg, characterized in that, Using 5-hydroxymaltol, maltol, N-methyltryptamine, adipic acid, and myristicin B1 as marker components, the test samples were detected by ultra-high performance liquid chromatography-ultraviolet detection / quadrupole-time-of-flight mass spectrometry. The content of the marker components determined whether the test sample was raw or bran-processed nutmeg.

2. The method according to claim 1, characterized in that, The sample to be tested is an aqueous extract of raw nutmeg or an aqueous extract of bran.

3. The method according to claim 1, characterized in that, The chromatographic conditions for the ultra-high performance liquid chromatography-ultraviolet detection / quadrupole-time-of-flight mass spectrometry are as follows: Chromatographic column: Octadecylsilane-bonded silica gel column; Mobile phase A is 0.05%~0.15%. v / v Formic acid aqueous solution, with acetonitrile as mobile phase B, is used for linear gradient elution. The procedure for the linear gradient elution is as follows: 0-22 minutes, 5%-42% B; 22-26 minutes, 42% B; 26-30 minutes, 42%-55% B; 30-37 minutes, 55%-80% B; 37-43 minutes, 80%-95% B; Flow rate: 0.25~0.35 mL / min; Column temperature: 35~45 ℃.

4. The method according to claim 3, characterized in that, The chromatographic column is selected from: CORTECS C18, with dimensions of 2.1 × 100 mm and a thickness of 1.6 μm; or CORTECS UPLC C18+, with dimensions of 2.1 × 100 mm and a thickness of 1.6 μm; or HSS T3, specifications: 2.1 × 100 mm, 1.8 μm; or BEH C8, with dimensions of 2.1 × 100 mm and a micrometer diameter of 1.7 μm; or BEH C18, with dimensions of 2.1 × 100 mm and a micrometer diameter of 1.7 μm; or CSH C18, with dimensions of 2.1 × 100 mm and a micrometer diameter of 1.7 μm; or CSH Phenyl-Hexyl, specifications: 2.1 × 100 mm, 1.7 μm; or CSH Cyano, with dimensions of 2.1 × 100 mm and 1.8 μm; or CSH Fluoro-Phenyl, with specifications of 2.1 × 100 mm and 1.7 μm; or Atlantis Premier BEH C18 AX, with dimensions of 2.1 × 100 mm and a thickness of 1.7 μm; or Zorbax SB-Aq, with dimensions of 2.1 × 100 mm and a thickness of 1.8 μm; or Zorbax Extend C18, with dimensions of 2.1 × 100 mm and a thickness of 1.8 μm; or Kinetex 1.7μ Biphenyl, with a specification of 2.1×100 mm and 1.7 μm; or Kinetex XB-C18, with dimensions of 2.1 × 100 mm and a micrometer diameter of 1.7 μm; or Luna Omega 1.6µ Polar C18, with dimensions of 2.1×100 mm and 1.6 μm.

5. The method according to claim 4, characterized in that, The chromatographic column was a Luna Omega Polar C18 with dimensions of 2.1 × 100 mm and a diameter of 1.6 μm.

6. The method according to claim 3, characterized in that, The mobile phase A is 0.1%. v / v Formic acid aqueous solution; and / or The flow rate is 0.3 mL / min; and / or The column temperature is 40 ℃.

7. The method according to any one of claims 3 to 6, characterized in that, The ultra-high performance liquid chromatography-ultraviolet detection / quadrupole-time-of-flight mass spectrometry (UHPLC-UV detection / quadrupole-time-of-flight mass spectrometry) uses positive ion mode ESI for mass spectrometry acquisition, with a scanning range of... m / z 100-800; Mass spectrometry conditions included: capillary voltage 4000 V, nozzle voltage 500 V, fragmentor voltage 380 V, drying gas temperature 200 °C, drying gas flow rate 12 L / min, nebulizer pressure 35 psi, sheath gas temperature 350 °C, and collision energy set to 5–30 V.

8. The application of the method according to any one of claims 1 to 7 in constructing a machine learning model for identifying raw nutmeg and its bran-processed products.

9. The application according to claim 8, characterized in that, The machine learning algorithms used in the machine learning model include support vector machine, partial least squares discriminant analysis, and random forest.

10. The application according to claim 9, characterized in that, The machine learning algorithm is either a support vector machine or a random forest.

Images