Fingerprint method for distinguishing different styles of tobacco-growing area soil by using soil characteristic metabolites

Abstract

The invention discloses a fingerprint spectrum method for distinguishing different styles of tobacco area soil by using soil characteristic metabolites. The method comprises the following steps: S1, collecting the soil of different aroma type tobacco leaf production areas as samples; s2, extracting non-volatile metabolites in the soil sample; s3, performing chromatographic analysis on each non-volatile metabolite; s4, taking a chromatogram with the largest number of chromatographic peaks obtained after UPLC on-machine analysis of the sample as a reference spectrum, and determining that common peaks of the chromatograms of the collected sample are fingerprint characteristic peaks of the fingerprint spectrum of the non-volatile metabolites in the soil sample through Mark peak matching; s5, analyzing and judging the significant difference level of the non-volatile metabolites in the soil samples collected in the tobacco leaf producing areas with different aroma types; s6, analyzing and screening different metabolites influencing classification in the soil of the tobacco leaf producing areas with different aroma types; s7, comparing the characteristic differences of metabolites in the soil of the tobacco leaf producing areas with different aroma types; and S8, verifying the applicability of the differential metabolites to the classification of the tobacco leaf production areas with different aroma types by utilizing a clustering and discriminant equation, and predicting the tobacco leaf production area to which the unknown soil sample belongs.

Description

Technical Field

[0001] The present invention relates to a fingerprint method for distinguishing soils of different style tobacco-growing areas by soil characteristic metabolites, and belongs to the technical field of soil metabolite analysis. Background Art

[0002] Tobacco leaf raw materials are an important basis for stabilizing the quality and style of cigarette products. Tobacco leaf raw materials from different sources are affected by factors such as ecological environment and genetic characteristics, resulting in different aroma styles. Guizhou is located in mountainous areas at low latitudes, with the terrain increasing from east to west, and obvious changes in various ecological factors. There are differences in the quality, chemical composition and aroma-producing substances of flue-cured tobacco in each ecological area, resulting in different aroma styles of tobacco leaves. As an important component of ecological factors, soil conditions are closely related to the formation of tobacco leaf quality and style characteristics.

[0003] Soil organic matter (SOM) is the basic substance supplied by the ecosystem, providing the energy required for metabolism for the exertion of soil functions. Affected by spatial and temporal changes, its material composition has regional differences. Luo Yi et al. found in their research (Luo Yi, Qin Song, Hu Gang, et al. Content characteristics of active organic matter and organic nitrogen in tobacco-growing soils of three aroma types in Guizhou [J]. Southwest China Journal of Agricultural Sciences, 2017, 30(8): 1833-1837.) that the total amount of active organic matter in soils of different aroma types of tobacco is in the order of light aroma type > slightly light aroma type > intermediate aroma type, and there are significant differences in the content of the same active organic matter among soils of different aroma types of tobacco. The research by Yan et al. showed (Yan S, Niu Z, Yan H, et al. Influence of soil organic carbon on the aroma of tobacco leaves and the structure of microbial communities [J]. Current Microbiology, 2020, 77: 931-942.) that soil active organic matter significantly increased the content of neutral aroma substances in tobacco leaves by affecting the biosynthesis transcription processes of sesquiterpenoids and chlorophyll, and believed that the composition and content distribution of active organic matter might be important factors in the formation of tobacco leaf aroma styles. Soil non-volatile metabolites (complex small molecule soil metabolite mixtures), as part of active SOM, include various chemical substances such as amino acids, polypeptides, lipids, carbohydrates and a large number of single macromolecules, which are the comprehensive expression results of metagenome, metatranscriptome and metaproteome at the biological level and are the manifestation of the material cycle of biological communities. Therefore, studying the composition and distribution characteristics of non-volatile metabolites in soil can analyze the role played by small molecule organic substances in soil in the biochemical system related to "regulating, managing and supporting ecosystem supply", and is the key to further clarifying the complex relationship between organisms and the ecological environment.

[0004] In recent years, GC-MS or LC-MS-based environmental metabolomics techniques have often been used for non-targeted / targeted analysis of the composition of small molecule metabolites in soil matrices, to analyze the relationship between small molecule organic compounds in soil and soil nutrient quality, and the metabolic interactions between plants and microorganisms. However, this technique often has problems such as expensive instrumentation, the need to identify substances, high requirements for the professional level of personnel, and high costs.

[0005] Fingerprint technology can find homogenous common compounds under a fuzzy attribute background, and combined with chemometrics for differential comparison, it can achieve origin tracing of complex systems such as drugs and foods, and differential discrimination of quality evaluation, with characteristics such as simple methods and wide applicability. Wang Yan et al. (Wang Yan, Fan Jiali, Zhang Minmin, et al. Analysis of Volatile Components of American Ginseng from Different Origins Based on Gas Chromatography-Ion Mobility Spectrometry [J]. Journal of Instrumental Analysis, 2023, 42(8): 976-983.) established a fingerprint of volatile components in American ginseng from different domestic and foreign origins using GC-IMS, and combined with PLS-DA to screen differential chemical components for the identification of American ginseng origins. Zhang Yashu et al. (Zhang Yashu, Huang Meiting, Lü Shijun, et al. Method for Traceability of Brown Sugar Origin Based on HPLC-DAD Fingerprint Combined with Chemical Pattern Recognition [J]. Food Research and Development, 2024, 45(8): 149-158.) established a fingerprint of non-volatile components of brown sugar from different origins based on HPLC-DAD, and combined with common patterns, similarity evaluation, and chemometrics to achieve the traceability of brown sugar origin and quality control. In tobacco research, it has also been used for origin tracing of different cigarette brands, screening of characteristic aroma components of different flue-cured tobacco types, and discrimination of different tobacco types. However, there are few reports on using fingerprint technology of non-volatile metabolites in soil to analyze the influence of soil ecological mechanisms on the formation of different flue-cured tobacco styles and conduct soil traceability. Summary of the Invention

[0006] Based on the above, the present invention provides a fingerprint method for distinguishing soils of different flue-cured tobacco areas by soil characteristic metabolites to overcome the deficiencies of the prior art.

[0007] The technical solution of the present invention is: a fingerprint method for distinguishing soils of different flue-cured tobacco areas by soil characteristic metabolites, including:

[0008] S1 Collection: Collect soils from different flue-cured tobacco leaf production areas as samples;

[0009] S2 Extraction: Extract non-volatile metabolites from the soil samples;

[0010] S3 Chromatographic analysis: Perform chromatographic analysis on each non-volatile metabolite to obtain chromatogram data;

[0011] S4 Fingerprint map establishment: Using the chromatogram with the largest number of chromatographic peaks obtained after the UPLC analysis of soil samples as the reference chromatogram, through Mark peak matching, the common peaks of the collected sample chromatograms were determined as the fingerprint characteristic peaks of the non-volatile metabolite fingerprint map of soil samples;

[0012] S5 Discrimination of differences in tobacco leaf production areas of different fragrance types: Using the common peak area as the independent variable and the tobacco leaf production area of the fragrance type to which the soil sample belongs as the dependent variable, the soil samples were marked according to their production areas, and the significant level of differences in non-volatile metabolites in soil samples collected from different tobacco leaf production areas of different fragrance types was determined through PCA analysis;

[0013] S6 Screening of differential metabolites: Using PLS-DA analysis combined with one-way ANOVA to screen the differential metabolites affecting classification in soil samples from different tobacco leaf production areas of different fragrance types;

[0014] S7 Characteristic analysis of differential metabolites: Comparing the characteristic differences of metabolites in soils from different tobacco leaf production areas of different fragrance types based on the changing trends of the mean values and sum values of the peak areas of each differential metabolite;

[0015] S8 Verification of the applicability of differential metabolites: Using HCA and Fisher discriminant equations to verify the applicability of differential metabolites to the classification of different tobacco leaf production areas and predict the tobacco leaf production area to which an unknown soil sample belongs;

[0016] Preferably, in step S1, the different tobacco leaf production areas of different fragrance types include Honey-Sweet Area I, II and Fresh-Sweet Area.

[0017] Preferably, in step S2, the extraction method is: adding acetonitrile to the soil sample for homogenized extraction, centrifuging, taking the supernatant and drying it with nitrogen, and then redissolving it with a methanol-aqueous solution.

[0018] Preferably, in step S2, the conditions for homogenized extraction are: frequency 60 Hz, time 5 min / time, and homogenized extraction 3 times.

[0019] Preferably, in step S3, the conditions for chromatographic analysis are: mobile phase A: water, B: acetonitrile; injection volume: 5 μL; column temperature: 30 °C; flow rate: 0.3 mL / min; gradient elution program: 0 - 11 min, 2% - 90% B; 11 - 14 min, 90% - 100% B; 14.1 min 2% B; running time: 17 min; detection wavelength: 230 nm; using chromatographic software to integrate chromatographic peaks, algorithm: traditional; peak width 6 s; minimum height 100 μV; minimum area 300 μV*s.

[0020] Preferably, in step S4, the chromatogram data of the soil sample is imported into the "Similarity Evaluation System for Chromatographic Fingerprints of Traditional Chinese Medicine" software, and Mark peak matching is performed to determine that the common peaks in the chromatogram of the collected sample are the fingerprint characteristic peaks of the non-volatile metabolite fingerprint of the soil sample, and the UPLC fingerprint of the non-volatile metabolites in the soil of different tobacco leaf producing areas with different fragrance types is established.

[0021] Preferably, in step S6, according to the VIP value > 1.0 and the p value < 0.05, the differential metabolites affecting classification in the soil of different tobacco leaf producing areas with different fragrance types are screened out.

[0022] Preferably, in step S7, the total mean of the differential metabolites shows the trend of Honey Sweet II > Honey Sweet I > Refreshing Sweet area, and the mean of each differential metabolite shows the trend that most of the Honey Sweet areas are higher than the Refreshing Sweet area.

[0023] Preferably, in step S8, the within-group sum of squares clustering analysis method is used for clustering, and the Fisher discriminant function model is used for the discriminant equation, which can accurately classify 4 external verification samples into their respective tobacco leaf producing areas.

[0024] Beneficial effects of the present invention: By utilizing the characteristics of fingerprint spectra to achieve differential analysis by finding common compounds under a fuzzy attribute background, and combining the discriminant technology of chemical pattern recognition, the present invention realizes the differential comparison of the content distribution of non-volatile metabolites in the soil of different tobacco leaf producing areas with different fragrance types on a simple liquid chromatograph, and screens out differential metabolites, providing a new theoretical basis for the identification of unknown soil samples in different tobacco leaf fragrance areas, the directional selection of characteristic tobacco leaf areas, and the relationship between the formation of tobacco leaf styles and soil ecology. Description of the Drawings

[0025] Figure 1 UPLC common peak pattern diagram (a) and reference fingerprint spectrum (b);

[0026] Figure 2 PCA score diagram of soil samples (a) and cross-validation of PLS-DA model (b);

[0027] Figure 3 VIP score diagram of soil metabolites;

[0028] Figure 4 Differential metabolite peak area distribution diagram;

[0029] Figure 5 External verification pedigree diagram of soil samples;

[0030] Figure 6 PCA score diagram of soil samples from different tobacco leaf areas;

[0031] Note: MTⅠ refers to Honey Sweet area Ⅰ, MTⅡ refers to Honey Sweet area Ⅱ, QT refers to Refreshing Sweet area, the same below.

[0032] Figure 7 Heatmap of the correlation between the contents of differential metabolites and physicochemical indexes; Note: * indicates significant correlation (p<0.05); ** indicates extremely significant correlation (p<0.01);

[0033] Figure 8 UPLC chromatograms of QC samples after extraction under different homogenization conditions; Note: a: 30Hz, 3min / time; b: 60Hz, 1min / time; c: 60Hz, 3min / time; d: 60Hz, 5min / time; e: 60Hz, 15min / time;

[0034] Figure 9 Flow chart of the fingerprint method. Detailed implementation manners

[0035] To make the above objects, features and advantages of the present invention more obvious and understandable, the following will describe the detailed implementation manners of the present invention with reference to the accompanying drawings. Many specific details are set forth in the following description in order to fully understand the present invention. However, the present invention can be implemented in many other ways different from those described herein, and those skilled in the art can make similar improvements without departing from the connotation of the present invention. Therefore, the present invention is not limited by the specific implementations disclosed below.

[0036] Example 1: A fingerprint method for distinguishing soils of different style tobacco-growing areas by soil characteristic metabolites

[0037] This method includes the following steps:

[0038] S1 Sampling: Collect soils from different flavor tobacco leaf production areas as soil samples.

[0039] Collect 18 soil samples (numbered S1-S18) from typical tobacco-growing areas in Guizhou, and divide them into Honey-sweet area Ⅰ (Zunyi tobacco-growing area with typical honey-sweet flavor), Honey-sweet area Ⅱ (Tongren tobacco-growing area) and Light-sweet area (Xingyi tobacco-growing area) according to the sampling locations, as shown in Table 1. The above all adopt the five-point sampling method, and take 1 kg of mixed soil from the 0-20 cm cultivation layer.

[0040] Table 1 Origin information of 18 soil samples from different tobacco-growing areas

[0041]

[0042]

[0043] S2 Extraction: Extract non-volatile metabolites from the soil samples.

[0044] Weigh 0.500 g of dry weight soil sample in a 2 mL centrifuge tube, add 1 mL of acetonitrile, and homogenize and extract 3 times at 60 Hz, each time for 5 min. After homogenization, centrifuge at 12000 rpm for 5 min, take 600 μL of supernatant and dry it with nitrogen in a 2 mL centrifuge tube, and re-dissolve it with 100 μL of 50% methanol-water (V / V) solution.

[0045] S3 Chromatographic analysis: Perform chromatographic analysis on each non-volatile metabolite to obtain chromatographic data.

[0046] The metabolites after reconstitution were analyzed by UPLC. Chromatographic conditions: mobile phase A: water, B: acetonitrile; injection volume: 5 μL; column temperature: 30 °C; flow rate: 0.3 mL / min; gradient elution program: 0-11 min, 2%-90% B; 11-14 min, 90%-100% B; 14.1 min 2% B; run time: 17 min; detection wavelength: 230 nm. UPLC Empower 3 chromatography software was used to integrate the chromatographic peaks (algorithm: traditional; peak width 6 s; minimum height 100 μV; minimum area 300 μV*s). By comparing the UPLC chromatogram of the blank sample, it was determined that the peak time of non-volatile metabolites in the soil in the sample was mainly distributed in 4-12 min, and all peaks that could be integrated within this range were calibrated as target peaks.

[0047] S4 fingerprint establishment: The chromatogram with the largest number of chromatographic peaks obtained after ultra-high performance liquid chromatography (UPLC) analysis of soil samples is used as the reference spectrum. After Mark peak matching, the common peaks in the chromatograms of the collected samples are determined as the fingerprint characteristic peaks of the fingerprint spectrum of non-volatile metabolites in the soil samples.

[0048] UPLC chromatographic analysis was performed on soils S1 to S18. According to the peak distribution, the chromatogram data of 18 soil samples for 4 to 12 minutes were imported into the "Similarity Evaluation System for Chromatographic Fingerprints of Traditional Chinese Medicine" software for characteristic peak analysis. The S10 sample spectrum with the largest number of chromatographic peaks was selected as the reference spectrum. After Mark peak matching, 30 common peaks were identified as characteristic peaks of the fingerprint spectrum of non-volatile metabolites in soil samples from typical tobacco-producing areas in Guizhou. The average method combined with multi-point correction (setting the correction time window width to 0.1) was used to generate a control fingerprint spectrum ( Figure 1 ). The chromatographic peaks represent different chemical components, and the size of the chromatographic peak area represents the content of each component.

[0049] The overall correlation between different fingerprint spectra was evaluated using the similarity of fingerprint spectra. The results showed (Table 2) that the similarity between S1 - S18 and the control fingerprint spectrum was between 0.809 and 0.956, indicating that the fingerprint spectra of soil samples from each tobacco-growing area had a high similarity with the control fingerprint spectrum, but there were certain differences. Thus, it can be seen that the non-volatile metabolite compositions in the soil samples for different flue-cured tobacco types were similar, but their content distributions were different, which was in line with the principle of the fingerprint spectrum evaluation results.

[0050] Table 2 Calculation results of the similarity of 18 soil samples

[0051] Number Similarity Number Similarity S1 0.956 S10 0.916 S2 0.934 S11 0.818 S3 0.937 S12 0.936 S4 0.942 S13 0.809 S5 0.918 S14 0.845 S6 0.918 S15 0.922 S7 0.933 S16 0.893 S8 0.841 S17 0.883 S9 0.907 S18 0.922

[0052] Differential discrimination of different flue-cured tobacco types in different producing areas: Using the common peak area as the independent variable and the flue-cured tobacco type producing area to which the soil sample belongs as the dependent variable, the soil samples were marked according to their producing areas, and the significant level of the differences in non-volatile metabolites in the soil samples collected from different flue-cured tobacco type producing areas was determined through PCA analysis.

[0053] PCA analysis was performed using the 30 common peak areas of the S1 - S18 soil samples as variables. As Figure 2 shown in a, the cumulative variance contribution rate of the first two principal components was 75% (>60%), indicating that the two principal components extracted could represent the overall characteristics of the fingerprint spectra of the 18 soil samples. The soil samples from Honey-sweet Area Ⅰ, Honey-sweet Area Ⅱ, and Light-sweet Area were clustered into three categories, indicating that there were significant differences in the metabolite content distributions among the Honey-sweet and Light-sweet areas, and between Honey-sweet Area I and Honey-sweet Area Ⅱ.

[0054] Screening of differential metabolites: PLS-DA analysis combined with one-way ANOVA was used to screen the differential metabolites affecting classification in the soils of different flue-cured tobacco type producing areas.

[0055] A PLS-DA model was established to screen the differential metabolites affecting the classification of samples from three tobacco-growing areas. From the model analysis and verification parameters, it can be seen that R X 2 (X matrix interpretation rate parameter) = 0.599, R Y 2 (model stability parameter) = 0.954, Q 2 (prediction ability parameter) = 0.919 (both >0.5), indicating that the model had high stability and prediction rate. The classification Y matrix variable was set to be randomly arranged 200 times for permutation test, and the permutation test graph ( Figure 2 b) was obtained. The results showed that the intercept of the fitting line on the Y-axis of R 2 was 0.283 (<0.3), indicating that the model was reliable, and the intercept of the fitting line on the Y-axis of Q 2 was -0.318 (<0.05), indicating that the established model did not have overfitting phenomenon.

[0056] According to VIP value>1.0 and p value<0.05, 17 metabolites with significant differences in soils from three different tobacco-growing areas were screened out ( Figure 3 ), which are peaks 3, 4, 6, 9, 10, 13, 14, 16, 18, 19, 21, 22, 23, 25, 28, 29, and 30 respectively.

[0057] S7 Analysis of differential metabolite characteristics: The characteristic differences of metabolites in the soil of different tobacco production areas were compared based on the changing trends of the mean and sum values ​​of the peak areas of each differential metabolite.

[0058] The peak area distribution characteristics of differential metabolites in soils from different tobacco growing areas are shown in Figure 4 . The results showed that the total content of differential metabolites in the honey sweet area Ⅱ was significantly higher than that in other tobacco areas, specifically, honey sweet area Ⅱ> honey sweet area Ⅰ> clear sweet area. The content of differential metabolites in the soil of the honey sweet area was mostly higher than that of the clear sweet area, among which peaks 4, 6, 10, 13, 14, 18, 25, 28, and 29 were significantly higher in the honey sweet area samples than in the clear sweet area, and peaks 3 and 9 were significantly lower than those in the clear sweet area, indicating that the content of non-volatile metabolites in the soil samples of the honey sweet area was richer. The content of each differential metabolite between the honey sweet area Ⅰ and the honey sweet area Ⅱ was significantly different, among which the content of peaks 10, 13, 16, 21, 22, 23 and 30 in the honey sweet area I samples was significantly higher than that in the honey sweet area Ⅱ, and the content of peaks 4, 6, 9, 14, 18, 19, 25, 28 and 29 were significantly lower than those in the honey sweet area Ⅱ, and only the content of peak 3 was not significantly different.

[0059] S8 Verification of the applicability of differential metabolites: HCA and Fisher discriminant equation were used to verify the applicability of differential metabolites to the classification of different aroma tobacco producing areas, and to predict the tobacco producing area to which the unknown soil samples belonged.

[0060] ① Cluster analysis verification

[0061] Four soil samples (numbered S19-S22) were randomly collected from Houchang, Weng'an, Qiannan Prefecture, Shangtang, Huangping, Qiandongnan Prefecture, Yaxi, Bozhou, Zunyi City, and Heishiyan District, Weining, Bijie City, for external sample traceability verification. Cluster analysis was performed using the sum of squares cluster analysis method with 17 differential metabolite peak areas of 18 soil samples (S1-S18) and 4 external samples as variables. The results showed that ( Figure 5 ), the four external samples were accurately divided into zones, S19 (Weng'an Houchang, Qiannan Prefecture), S20 (Shangtang, Huangping, Qiandongnan Prefecture) and S21 (Yaxi, Bozhou, Zunyi City) were classified into honey sweet zone I, and S22 (Heishi, Weining, Bijie City) was classified into fresh sweet zone. This indicates that the 17 differential metabolites obtained by fingerprint spectrum combined with discriminant analysis method are well representative in reflecting the soil differences of different aroma tobacco areas.

[0062] ② Fisher discriminant function model verification (X: Peak areas of 17 characteristic metabolites, Y1: Honey-sweet area I;

[0063] Y2: Honey-sweet area II, Y3: Light-sweet area, and the maximum value is the corresponding tobacco leaf fragrance type of the soil)

[0064] Compared with cluster analysis, Fisher discriminant analysis can perform quantitative discrimination. It is a linear analysis method for classification that can minimize the differences within the same class of samples and maximize the differences between different classes of sample points, thereby effectively improving the discrimination efficiency. Using the Fisher quantitative discriminant function model (Y1: Honey-sweet area I; Y2: Honey-sweet area II; Y3: Light-sweet area), the discrimination effect of the 17 differential metabolites screened out in the origin tracing of soil samples was further verified. Taking the peak areas of 17 differential metabolites in 18 soil samples as characteristic indicators, the original discrimination analysis and cross-validation of the samples were carried out; a total of 4 soil samples from the honey-sweet area I and the light-sweet area were selected as external samples for model verification. The discriminant analysis results show (Table 3) that the original discrimination accuracy rate is 100%, the average accuracy rate of cross-validation is 95.5%, and the external verification accuracy rate is 100%. Compared with cluster analysis, the discriminant function model can perform quantitative discrimination.

[0065] Y1 = 43.811X3 - 11.177X4 - 46.433X6 - 20.953X9 + 18.292X 10 - 17.462X 13 - 50.305X 14 + 6.327X 16 + 3.618X 18 - 40.839X 19 + 27.402X 21 - 11.695X 22 - 10.305X 23 - 59.721X 25 - 32.981X 28 + 32.284X 29 - 6.127X 30 - 88.429

[0066] Y2 = - 344.321X3 - 185.526X4 + 92.928X6 + 204.429X9 + 149.402X 10 + 171.361X 13 + 292.261X 14 - 150.172X 16 + 10.565X 18 + 270.771X 19 - 132.852X 21 + 54.1X22 -52.712X 23 +134.148X 25 +478.65X 28 -135.442X 29 -138.132X 30 -787.894

[0067] Y3 = 238.804X3 + 173.393X4 - 19.953X6 - 148.286X9 - 151.577X 10 -124.429X 13 -185.832X 14 +120.585X 16 -13.708X 18 -179.583X 19 +78.642X 21 -31.335X 22 +58.43X 23 -38.2X 25 -367.867X 28 +74.586X 29 +110.522X 30 -342.565

[0068] Table 3 Fisher Discriminant Analysis of Soil Sample Classification in Different Tobacco-growing Areas

[0069]

[0070] Note: The original discriminant accuracy rate is 100%.

[0071] The research process of the present invention will be described in detail below:

[0072] I. The applicant first analyzed the differences in the physical and chemical properties of soil samples from different flavor-type tobacco leaf production areas

[0073] Determine the physical and chemical indexes of the soil samples S1 - S18 in Example 1. The total nitrogen (TK), total phosphorus (TP), total potassium (TK), hydrolyzable nitrogen (AHN), available phosphorus (AP), available potassium (AK), organic matter (OM) content, pH value and water content in the soil samples were determined by the standard methods of LY / T 1228 - 2015, NY / T88 - 1988, NY / T 87 - 1988, NY / T 1121.25 - 2012, NY / T 889 - 2004, NY / T 1121.6 - 2006, NY / T 1377 - 2007, LY / T 1213 - 1999.

[0074] The physical and chemical properties of 18 soil samples from different tobacco-growing areas are shown in Table 4. There were no significant differences in the contents of TP, TK, AP, and OM among the soil samples from different tobacco-growing areas. There were significant differences in the contents of TN and AHN between the soil samples from Honey-Sweet Area II and Sweet Area. There were significant differences in the content of AK between the soil samples from Honey-Sweet Area I and Sweet Area. There were significant differences in the pH value among the soil samples from Honey-Sweet Area I, Honey-Sweet Area II, and Sweet Area.

[0075] Table 4 Multiple comparison analysis of the physical and chemical properties of soil samples from different tobacco-growing areas

[0076]

[0077] Note: The data are presented in the form of mean ± standard deviation and are based on the results of one-way ANOVA. Duncan's multiple comparison was used. Different letters in the same column indicate significant differences at the p = 0.05 level.

[0078] PCA analysis was performed with the physical and chemical property indexes of S1-S18 soils as variables. The results showed that Figure 6 ), the first two principal components could represent the overall physical and chemical characteristics of soil samples from different tobacco-growing areas, but could not distinguish the three tobacco-growing areas. There was a large overlap between Sweet Area and Honey-Sweet Area II, indicating that the conventional physical and chemical property indexes of soil could not characterize the changes in soil ecological factors in different tobacco-growing areas.

[0079] The correlation analysis results between soil differential metabolites and physical and chemical property indexes showed that Figure 7 ), the differential metabolites were only significantly negatively correlated with the pH value and not significantly correlated with other indexes.

[0080] II. Parameter optimization of the present invention

[0081] 1. Determination of chromatographic conditions

[0082] The UPLC BEH C18 chromatographic column was selected to separate the target compounds, and the conditions such as the composition of the mobile phase, column temperature, injection volume, and detection wavelength were optimized. The separation effects of metabolites in soil samples were compared under the conditions of isocratic elution with 90% acetonitrile / water (V / V) and gradient elution with 10%-90% acetonitrile-water (running time 10 min). The results showed that gradient elution could achieve good separation of target compounds with different polarities and optimized the best gradient elution conditions; the column temperature had little effect on the separation effect, so the commonly used 30 °C was selected as the separation condition; when the injection volume was 5 μL, the solvent effect of the chromatographic peak was small and the response intensity of the target peak was high; by using PDA full-wavelength scanning, the best wavelength with an absorption wavelength of 230 nm was determined, the baseline noise had little effect, and the number of target peaks was large.

[0083] 2. Optimization of extraction conditions

[0084] In the selection of extraction solvents, three different polar organic solvents commonly used for the extraction of non-volatile metabolites in soil (acetonitrile, ethanol, and methanol) were compared. The results showed that the extraction effects of ethanol and methanol were similar, and acetonitrile obtained the largest number of UPLC chromatographic peaks and the highest peak response values. Further comparison of the extraction effects of acetonitrile with 80%, 60%, and 50% aqueous acetonitrile solutions found that as the polarity of the extraction solvent increased, the peak response value gradually decreased, indicating that the polarity of the non-volatile organic compounds extracted from the soil was relatively weak. Therefore, acetonitrile was selected as the extractant.

[0085] In terms of extraction methods, the extraction effects of three common extraction methods, namely ultrasonic extraction (ultrasonic extraction for 35 min at 50 Hz, with the temperature not exceeding 30 °C), homogenization (homogenization 3 times at a frequency of 60 Hz and 4 °C, 3 min each time), and vortex mixing (mixing for 5 min at 2500 rpm), were compared. The results showed that the extraction effects of the three extraction methods were similar. However, ultrasonic extraction had a long extraction time, and vortex mixing extraction had the disadvantages of less extraction solution, insufficient contact with the sample, and poor peak area stability. Therefore, the homogenization method was selected as the best extraction method. The homogenization frequency (60 Hz and 30 Hz) and time (1 min, 3 min, 5 min, and 15 min each time, homogenization 3 times) were optimized. The results showed that when the extraction time was the same, the UPLC peak response value increased with the increase of the extraction frequency ( Figure 8 ). The peak response value increased with the increase of the homogenization time. However, when the homogenization time increased to 15 min each time, the extraction effect was not significantly different from that at 5 min each time, indicating that the extraction was complete at 5 min each time ( Figure 8 ). Therefore, the optimal homogenization conditions selected were a frequency of 60 Hz, a time of 5 min each time, and homogenization extraction 3 times.

[0086] To improve the detection sensitivity of the method, the homogenized extract was concentrated by nitrogen blowing. The results showed that the number of detected peaks and the response values increased significantly, and the composition of non-volatile metabolites was not changed.

[0087] III. Methodological verification

[0088] According to the method of the present invention, 5 parallel soil samples were measured daily for 3 consecutive days, and the intra-day and inter-day RSD% of the retention time (T) and peak area (S) of the target chromatographic peaks were calculated. The results showed that the intra-day precision was T RSD％ <1.71%, S RSD％ <2.21%; the inter-day precision was T RSD％ <2.14%, S RSD％ <2.75%, indicating that the precision and stability of this method were good.

[0089] The above-described embodiments merely represent several implementation manners of the present invention. The description thereof is relatively specific and detailed, but it should not be construed as a limitation to the scope of the patent for the present invention. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present invention, several modifications and improvements can still be made, and these all fall within the protection scope of the present invention. Therefore, the protection scope of the patent for the present invention shall be subject to the appended claims.

Claims

1. A fingerprint method for distinguishing soils from different tobacco-growing areas by soil characteristic metabolites, characterized in that: include: S1 Collection: Collect soil from tobacco production areas with different aroma types as samples; S2 extraction: extracting non-volatile metabolites from soil samples; S3 Chromatographic analysis: each non-volatile metabolite is subjected to chromatographic analysis to obtain chromatogram data; S4 fingerprint spectrum establishment: the chromatogram with the largest number of chromatographic peaks obtained after the ultra-high performance liquid chromatography analysis of the soil sample is used as the reference spectrum, and the common peaks of the chromatogram of the collected sample are determined as the fingerprint characteristic peaks of the fingerprint spectrum of the non-volatile metabolites in the soil sample through Mark peak matching; S5 Differential discrimination of tobacco production areas with different aroma types: Taking the common peak area as the independent variable and the tobacco production area to which the soil sample belongs as the dependent variable, the soil samples are marked according to the production area they belong to, and the significant level of differences in non-volatile metabolites in soil samples collected from tobacco production areas with different aroma types is determined by PCA analysis; S6 Screening of differential metabolites: PLS-DA analysis combined with one-way ANOVA was used to screen differential metabolites that affect classification in soil samples from tobacco production areas with different aroma types; S7 Differential metabolite characteristic analysis: The characteristic differences of metabolites in soils from tobacco production areas with different aroma types were compared based on the changing trends of the mean and sum values ​​of the peak areas of each differential metabolite; S8 Verification of the applicability of differential metabolites: Clustering and discriminant equations were used to verify the applicability of differential metabolites to the classification of different flavor tobacco producing areas, and to predict the tobacco producing area to which the unknown soil samples belonged.

2. The fingerprint method for distinguishing soils from different tobacco-producing areas by soil characteristic metabolites according to claim 1 is characterized in that: In step S1, the different flavor tobacco producing areas include sweet areas I, II and light sweet areas.

3. The fingerprint method for distinguishing soils from different tobacco-producing areas by soil characteristic metabolites according to claim 1 is characterized in that: In step S2, the extraction method is: adding acetonitrile to the soil sample for homogenous extraction, centrifuging, taking the supernatant and blowing it dry with nitrogen, and then re-dissolving it with a methanol-water solution.

4. The fingerprint method for distinguishing soils from different tobacco-producing areas by soil characteristic metabolites according to claim 3 is characterized in that: The conditions of the homogenous extraction are: frequency 60 Hz, time 5 min / time, and homogenous extraction 3 times.

5. The fingerprint method for distinguishing soils from different tobacco-growing areas by soil characteristic metabolites according to claim 1, characterized in that: In step S3, the conditions for chromatographic analysis are: mobile phase A: water, B: acetonitrile; injection volume: 5 μL; column temperature: 30°C; flow rate: 0.3 mL / min; gradient elution program: 0-11 min, 2%-90% B; 11-14 min, 90%-100% B; 14.1 min 2% B; running time: 17 min; detection wavelength: 230 nm; chromatographic peaks are integrated using chromatography software, algorithm: traditional; peak width 6 s; minimum height 100 μV; minimum area 300 μV*s.

6. The fingerprint method for distinguishing soils from different tobacco-growing areas by soil characteristic metabolites according to claim 1, characterized in that: In step S4, the chromatogram data of the soil sample is imported into the "Chinese Medicine Chromatographic Fingerprint Similarity Evaluation System" software, and the Mark peak is matched to determine the common peaks of the collected sample chromatograms as fingerprint characteristic peaks, and establish UPLC fingerprints of non-volatile metabolites in soils of tobacco producing areas with different aroma types.

7. The fingerprint method for distinguishing soils from different tobacco-growing areas by soil characteristic metabolites according to claim 1, characterized in that: In step S6, based on VIP value>1.0 and p value<0.05, differential metabolites affecting classification in soils of tobacco production areas with different aroma types are screened out.

8. The fingerprint method of soil characteristic metabolites for distinguishing soils from different tobacco styles according to claim 1 is characterized in that: In step S7, the sum of the mean values ​​of differential metabolites in the soil samples showed a trend of honey sweet II>honey sweet I>fresh sweet area, and the mean values ​​of each differential metabolite showed a trend that most honey sweet areas were higher than fresh sweet areas.

9. The fingerprint method of soil characteristic metabolites for distinguishing soils from different tobacco styles according to claim 1 is characterized in that: In step S8, the clustering method used is the sum of squares clustering analysis method, and the discriminant equation used is the Fisher discriminant function model.

Images