Non-targeted mass spectrometry identification method for tobacco carbonyl components

Abstract

A non-targeted mass spectrometry identification method for tobacco carbonyl components, comprising the following steps: (1) preparing a tobacco sample extract liquid; (2) performing derivatization by using 2,4-dinitrophenylhydrazine (DNPH) to obtain a first tobacco sample; (3) replacing DNPH with DNPH-d3 for derivatization to obtain a second tobacco sample; (4) preparing a mixed tobacco sample; (5) preparing a blank sample; (6) performing UPLC-Orbitrap-HRMS analysis respectively on the first tobacco sample, the second tobacco sample, the mixed tobacco sample, and the blank sample to obtain raw LC-MS data; (7) processing the raw LC-MS data to obtain raw mass spectrum characteristic data; (8) performing multiple filtering on the raw mass spectrum characteristic data; and (9) performing structural annotation or identification on the finally reserved chromatographic peak to obtain tobacco carbonyl components. By means of multiple filtering of mass spectrum characteristic data, chromatography-mass spectrum information of noise or interference components in a raw data set is quickly eliminated, and composition information of aldehyde and ketone chemical components in a cigarette is obtained.

Description

A non-targeted mass spectrometry method for identifying tobacco carbonyl components Technical Field

[0001] The present invention belongs to the field of tobacco technology, and specifically relates to a non-targeted mass spectrometry identification method for tobacco carbonyl components. The tobacco carbonyl components can be obtained by filtering mass spectrometry feature data and performing structural identification on retained chromatographic peaks. Background Art

[0002] Aldehydes and ketones are important carbonyl compounds that are ubiquitous in biological organisms and the human environment. Low-molecular-weight (carbon chain length <10) fatty aldehydes or ketones can react with biomolecules such as DNA, proteins, and enzymes, and are highly cytotoxic and genotoxic. Formaldehyde, acetaldehyde, and crotonaldehyde are classified as Group 1, Group 2B, and Group 3 human carcinogens, respectively, by the International Agency for Research on Cancer (IARC). Acrolein is listed on the "List of Hazardous Air Pollutants" by the US Environmental Protection Agency (US EPA); formaldehyde, acetaldehyde, acetone, acrolein, propionaldehyde, crotonaldehyde, 2-butanone, and butyraldehyde are all included on the Hoffmann list.

[0003] Aldehydes and ketones are the most numerous flavor components in tobacco, with over 500 identified in tobacco smoke. Given their significant impact on flavor and safety, comprehensive qualitative and quantitative analysis of aldehydes and ketones in cigarette smoke is crucial. Researchers have conducted extensive research on the qualitative and quantitative analysis of aldehydes and ketones in various systems using either gas chromatography-mass spectrometry (GC-MS) or liquid chromatography-mass spectrometry (LC-MS). LC-MS, combining the powerful separation capabilities of LC with the high specificity and sensitivity of MS, has become one of the most prominent techniques for the qualitative and quantitative analysis of small molecules. Depending on the research objective, targeted or non-targeted data acquisition approaches can be used to analyze target compounds. Targeted LC-MS methods primarily target known target compounds, providing accurate quantitative information through external standardization and other methods. However, targeted approaches typically focus on a small number of target compounds for which chemical standards are available. Complex sample systems may contain many potentially important components, but the lack of suitable chemical standards precludes targeted analysis.

[0004] Compared to targeted analysis, non-targeted LC-MS methods focus on analyzing and detecting all possible targets in a sample system. High-resolution mass spectrometry (HRMS), represented by ion trap (Orbitrap) mass spectrometers or time-of-flight (TOF) mass spectrometers, has been widely used for the non-targeted identification and detection of unknown components due to its extremely high resolution and mass accuracy. LC-HRMS can provide rich data such as precise molecular weight, isotopic distribution, and multi-stage mass spectrometry. Currently, LC-HRMS has been widely used in metabolomics, environmental analysis, and food safety analysis. By acquiring a large number of chromatographic-mass spectrometric signals covering both known and unknown components, it provides the possibility of a comprehensive understanding of the chemical composition of experimental samples.

[0005] Small molecule aldehydes and ketones have diverse structures, large polarity differences, high volatility, low ionization efficiency, and a wide concentration range. However, achieving highly sensitive non-targeted analysis of them using LC-MS still faces many difficulties. Chemical derivatization (chemical labeling) is an important method for improving the LC behavior of the target compound and enhancing the sensitivity and specificity of MS detection. To date, researchers have developed a series of chemical derivatization reactions combined with LC-QqQ-MS or LC-HRMS for the detection of aldehydes and ketones using p-nitrophenylhydrazine, 2,4-dinitrophenylhydrazine, Girard reagent T, and ammonium acetate-phenanthracene as derivatization reagents. This shows that the use of derivatization reactions combined with LC-HRMS detection can significantly improve the performance of LC-MS technology in detecting small molecule aldehydes and ketone compounds.

[0006] However, the datasets generated by LC-HRMS experiments are extremely large and complex. Processing this complex raw data to rapidly identify and determine target components in complex matrices has always been a very difficult task. First, the complex raw data contains thousands of mass spectral features, including a particularly large amount of redundant chromatographic-mass spectrometric information. Even a blank sample containing no target compound may detect thousands of well-shaped chromatographic peaks. In addition, the concentration distribution of compounds in complex samples is very wide, and the mass spectrometric signals of low-abundance target compounds may be overwhelmed by background noise ions or signals from matrix interfering ions due to their low intensity. Therefore, it is difficult to intuitively extract the chromatographic-mass spectrometric information of low-abundance target components from the raw data, which increases the difficulty of their qualitative and quantitative analysis.

[0007] Data cleaning and mass spectral information filtering can help us efficiently identify and characterize target components from complex LC-HRMS datasets, especially those with very low abundance and trace amounts. Based on the chromatographic behavior and mass spectral characteristics of target compounds, different data cleaning methods and mass spectral information filtering strategies can be designed to efficiently filter interfering ions from the raw data and target the target components. For example, metabolomics analysis uses the coefficient of variation of chromatographic peaks (mass spectral features) in quality control samples to assess stability, thereby eliminating a large number of interfering chromatographic peaks detected. In drug metabolism analysis, by setting a specific mass defect window, the chromatographic-mass spectrometric signals of many interfering components can be filtered out, allowing the identification of similar drug metabolites with specific structures. In natural product analysis, neutral loss filtering and diagnostic fragment ion filtering can be used to rapidly identify chemical components with similar chemical structures. However, many currently proposed LC-MS filtering methods still set filters based on a single dimension of chromatographic or mass spectral information, resulting in low information filtering efficiency and excessive false positive results. In particular, there are currently no mass spectrometric filtering methods that incorporate chemical derivatization labeling information. Therefore, by comprehensively utilizing the statistical laws, chromatographic behavior, mass spectrometric characteristics and other information of the target components, a multi-dimensional chromatography-mass spectrometry data filtering method is designed, which is expected to improve the filtering specificity and accuracy of the LC-MS raw data.

[0008] Overall, aldehydes and ketones are important chemical components in tobacco, and their non-targeted identification and identification is of great significance. However, an efficient method to achieve this goal is still lacking. The identification and identification of aldehydes and ketones in complex sample systems faces the following difficulties: 1. Due to their small molecular weight and low ionization efficiency, LC-MS techniques cannot directly detect them with high sensitivity; 2. The content distribution of aldehydes and ketones is wide, requiring a wide dynamic range for their analysis and detection; 3. The raw LC-MS data contain a large amount of chromatographic-mass spectrometric information of interfering components, which overwhelms the chromatographic-mass spectrometric signals of target components, especially those with low abundance, making the identification and identification of aldehydes and ketones very difficult.

[0009] The present invention is proposed for this purpose.

[0010] Summary of the Invention

[0011] The present invention uses derivatization reactions to chemically label potential aldehyde and ketone compounds, optimizes and establishes LC-HRMS methods for highly sensitive detection of aldehyde and ketone derivatization products, and proposes a multidimensional data filtering technology based on statistical characteristics, chromatographic behavior, mass spectral characteristics and isotope labeling information. Through automated processing, efficient recognition and identification of aldehyde and ketone components in complex LC-MS data sets can be achieved.

[0012] The technical solutions of the present invention are as follows:

[0013] A non-targeted mass spectrometry identification method for tobacco carbonyl components comprises the following steps:

[0014] (1) Wet a tobacco sample with water, then add a certain amount of acetonitrile and perform oscillation extraction to obtain a tobacco sample extract;

[0015] (2) taking a certain amount of tobacco sample extract, adding a certain amount of 2,4-dinitrophenylhydrazine (DNPH) thereto for derivatization, and preparing a first tobacco sample (CCs-DNPH);

[0016] (3) using DNPH-d3 instead of DNPH, repeating step (2) to prepare a second isotope-labeled tobacco sample (CCs-DNPH-d3);

[0017] (4) taking the first tobacco sample and the second tobacco sample in a ratio of 1:1, mixing them thoroughly, and filtering them through a filter membrane to obtain a third tobacco sample;

[0018] (5) Prepare a blank sample according to the procedures of steps (1) and (2); except that no tobacco sample is infiltrated with water, the other procedures are the same as steps (1) and (2);

[0019] (6) performing UPLC-Orbitrap-HRMS analysis on the first tobacco sample, the second tobacco sample, the third tobacco sample, and the blank sample to obtain raw LC-MS data;

[0020] (7) Peak detection, peak alignment, and peak grouping are performed on the raw LC-MS data to obtain several raw mass spectrometric features of the coexistence of target components and interfering components, including m / z values, retention times, and peak intensities;

[0021] (8) performing multiple filtering on the original mass spectrum characteristic data obtained in (7);

[0022] (9) Structural annotation or identification of the chromatographic peaks finally retained can obtain tobacco carbonyl components.

[0023] Preferably, the conditions for UPLC-Orbitrap-HRMS analysis are:

[0024] UPLC-HRMS was performed using an instrument platform consisting of a Dionex U3000 UHPLC system and a Q–Exactive mass spectrometer in tandem;

[0025] UPLC conditions: Syncronis C18 column (2.1 mm × 100 mm, 1.7 μm); column temperature: 40°C; injection volume: 1 μL; mobile phase A: 0.1% formic acid in water; mobile phase B: acetonitrile; gradient elution program: 95% B at 0–1 min, 95%–60% B at 1–3 min, 60%–10% B at 3–10 min, 10% B at 10–18 min, 10% B at 18–19 min, 10%–95% B at 19–20 min, 95% B; flow rate: 0.2 mL / min;

[0026] Mass spectrometry conditions included a spray voltage of 3.7 kV, a sheath gas flow rate of 35 L / min, an auxiliary gas flow rate of 10 L / min, and a DL transfer line temperature of 350°C. Data acquisition was performed in Full MS-DDA mode with negative ionization. The mass-to-charge ratio scan range for the primary mass spectrometer was set to 100–1200 m / z at a resolution of 70,000. The mass-to-charge ratio scan for the secondary mass spectrometer was set to a resolution of 35,000, and the high-energy collision-induced dissociation voltage was 30 eV.

[0027] Preferably, multiple filtering of mass spectral characteristic data includes: filtering of mass spectral characteristic data based on statistical characteristics, filtering of mass spectral characteristic data based on mass loss, filtering of mass spectral characteristic data of paired chromatographic peaks based on derivatization reaction labels, and filtering of mass spectral characteristics based on multiple ion secondary mass spectrometry to obtain the final retained chromatographic peaks.

[0028] Preferably, the steps of filtering mass spectrum characteristic data based on statistical features are:

[0029] Calculate the coefficient of variation CV and fold change FC, the calculation formula is CV = Int SQC / Int MQC ×100%; where Int SQC Represents the standard deviation of the peak intensity of a specific chromatographic peak in the first tobacco sample, Int MQC represents the average value of the peak intensity;

[0030] FC=Int MQC / Int MBK , where Int MQC Represents the average peak intensity of a specific chromatographic peak in the first tobacco sample, Int MBK represent the average values ​​of peak intensities of specific chromatographic peaks in blank samples;

[0031] If CV≤30% and FC≥1.5, the corresponding chromatographic peak was retained.

[0032] Preferably, the steps of filtering mass spectrum characteristic data based on mass defect are:

[0033] The calculation formula is: MD = |MZ–ceiling(MZ)|, where MZ represents the exact mass, i.e., the m / z value, of the precursor ion of the specific chromatographic peak detected in the first tobacco sample, and ceiling(MZ) is the ceiling number of the exact mass, i.e., the nominal mass;

[0034] If 0.02≤MD≤0.3 and m / z>209, retain the corresponding chromatographic peak.

[0035] Preferably, the step of filtering the mass spectrum characteristic data of the paired chromatographic peaks labeled by the derivatization reaction is: filtering the pair of chromatographic peaks detected in the third tobacco sample at the same retention time: if |MZ–MZ d3 |=3.0186, and |IntP1–IntP2| / max(P1,P2)<30%, and |RT1–RT2|<2, then retain the corresponding chromatographic peak; among them, MZ and MZ d3 where IntP1 represents the exact mass, i.e., the m / z value, of the precursor ions of the non-isotope-labeled and isotope-d3-labeled derivatization products, respectively; IntP1 and IntP2 represent the peak areas of the first and second chromatographic peaks, respectively; and RT1 and RT2 represent the retention times of the first and second chromatographic peaks, respectively.

[0036] Preferably, the filtering method based on the secondary mass spectrometry multiple ion mass spectrometry feature is: if the secondary mass spectrum of the first tobacco sample can produce m / z 76.018, m / z 120.008, m / z 122.024, m / z 135.019 and m / z 181.012 fragment ions, the corresponding chromatographic peaks are retained.

[0037] Preferably, the structural annotation or identification of the finally retained chromatographic peaks adopts methods including: standard matching, database search, or fragmentation pattern analysis.

[0038] Preferably, step (7) performs peak detection, peak alignment, and peak grouping, and step (8) performs multiple filtering on the mass spectrum characteristic data, and step (9) performs structural annotation or identification on the chromatographic peaks finally retained, which are automatically performed by a software package. The present invention develops a data processing package (MSFiltering package) based on R language for performing peak detection, peak alignment, and peak grouping on raw LC-MS data to obtain mass spectrum characteristics of the coexistence of target components and interfering components, a multidimensional data filtering method based on statistical characteristics, mass loss, paired chromatographic peaks, and multiple ion information, and performs structural annotation or identification on the chromatographic peaks finally retained, etc., through automated processing, to achieve efficient identification and identification of aldehydes and ketones in complex LC-MS data sets.

[0039] Beneficial effects of the present invention:

[0040] 1. The method of the present invention uses multiple filters on mass spectral feature data to rapidly remove noise or interfering chromatographic-mass spectrometric information from the original dataset, effectively identifying chromatographic peaks and mass spectral features that truly belong to aldehydes and ketones from complex, non-targeted datasets. The chemical structures of the extracted chromatographic peaks and mass spectral features belonging to aldehydes and ketones are annotated using methods such as standard sample matching, database retrieval, and fragmentation pattern analysis to obtain compositional information for the aldehydes and ketones in cigarette or food samples.

[0041] 2. The present invention has developed a data processing package (MSFiltering package) for performing peak detection, peak alignment and peak grouping on raw LC-MS data to obtain the mass spectral characteristics of the coexistence of target components and interfering components. A multidimensional data filtering method based on statistical characteristics, mass loss, isotope labeling and multiple ion information of secondary mass spectrometry, as well as structural identification of the finally retained chromatographic peaks, is used for automated processing to achieve efficient recognition and identification of aldehydes and ketones in complex LC-MS data sets. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 is a schematic diagram of the number of characteristic peaks of the comparative blank sample, the number of characteristic peaks after filtering the mass spectrum characteristic data based on statistical characteristics, the number of characteristic peaks after filtering the mass spectrum characteristic data based on mass loss, the number of characteristic peaks after filtering the mass spectrum characteristic data of paired chromatographic peaks labeled by derivatization reactions, and the number of characteristic peaks after filtering based on the multiple ion mass spectrum characteristics of the secondary mass spectrometry.

[0043] Figure 2 is a schematic diagram of the number of characteristic peaks of the comparative mixed sample, the number of characteristic peaks after filtering the mass spectrum characteristic data based on statistical characteristics, the number of characteristic peaks after filtering the mass spectrum characteristic data based on mass loss, the number of characteristic peaks after filtering the mass spectrum characteristic data of paired chromatographic peaks labeled by derivatization reactions, and the number of characteristic peaks after filtering based on the multiple ion mass spectrum characteristics of the secondary mass spectrometry.

[0044] Figure 3 is a schematic diagram of the recall rate of the comparative mixed standard sample, as well as the recall rate after filtering the mass spectral feature data based on statistical features, the recall rate after filtering the mass spectral feature data based on mass loss, the recall rate after filtering the mass spectral feature data of paired chromatographic peaks labeled by derivatization reactions, and the recall rate after filtering based on the multiple ion mass spectral features of the secondary mass spectrometry.

[0045] Figure 4 is a schematic diagram of the number of characteristic peaks of the tobacco sample in the embodiment, and the number of characteristic peaks after filtering the mass spectrum characteristic data based on statistical characteristics, the number of characteristic peaks after filtering the mass spectrum characteristic data based on mass loss, the number of characteristic peaks after filtering the mass spectrum characteristic data of paired chromatographic peaks labeled by derivatization reaction, and the number of characteristic peaks after filtering based on the multiple ion mass spectrum characteristics of secondary mass spectrometry. DETAILED DESCRIPTION

[0046] The present invention will be further described in detail below through specific embodiments, but this should not be construed as limiting the scope of the present invention to the following examples. Various substitutions and modifications made according to common technical knowledge and customary means in the art without departing from the spirit of the present invention are intended to be included within the scope of the present invention.

[0047] Example: Non-targeted mass spectrometry identification method for tobacco carbonyl components, the steps are as follows:

[0048] (1) Wetting a tobacco raw material sample with water, adding a certain amount of acetonitrile and performing oscillation extraction to obtain a tobacco sample extract;

[0049] (2) taking a certain amount of sample extract, adding a certain amount of 2,4-dinitrophenylhydrazine (DNPH) thereto for derivatization, and preparing a first tobacco sample (CCs-DNPH);

[0050] (3) using DNPH-d3 instead of DNPH, repeating step (2) to prepare a second isotope-labeled tobacco sample (CCs-DNPH-d3);

[0051] (4) taking the first tobacco sample and the second tobacco sample in a ratio of 1:1, mixing them thoroughly, and filtering them through a filter membrane to obtain a third tobacco sample;

[0052] (5) preparing a blank sample without water infiltration of the tobacco raw material sample according to steps (1) and (2);

[0053] (6) performing UPLC-Orbitrap-HRMS analysis on the first tobacco sample, the second tobacco sample, the third tobacco sample, the blank sample, and the mixed sample to obtain raw LC-MS data;

[0054] (7) Peak detection, peak alignment, and peak grouping are performed on the raw LC-MS data to obtain the mass spectral characteristics of the coexistence of target components and interfering components, including m / z values, retention times, and peak intensities;

[0055] (8) Perform multiple filtering on mass spectrometry feature data;

[0056] (9) The structure of the finally retained chromatographic peak is identified to obtain tobacco carbonyl components.

[0057] The specific steps are:

[0058] Step 1: Cut the tobacco raw material into pieces of approximately 0.5 cm × 0.5 cm; take 1.0 g, accurate to 0.1 mg, and place it in a 100 mL stoppered Erlenmeyer flask. Add 5 mL of water. After the sample is completely soaked in water, accurately add 30 mL of acetonitrile and extract on an oscillator at 150 rpm for 30 min to obtain a sample extract.

[0059] Step 2: Accurately pipette 1.0 mL of tobacco sample extract into a 10 mL volumetric flask, add 4 mL of DNPH, and dilute to volume with acetonitrile. Shake well and incubate at room temperature for 30 minutes for derivatization to prepare the first tobacco sample (CCs-DNPH).

[0060] Step 3: Using DNPH-d3 instead of DNPH, repeat step 2 to prepare a second isotope-labeled tobacco sample (CCs-DNPH-d3);

[0061] Step 4: 1 mL of each of the first tobacco sample (CCs-DNPH) and the second tobacco sample (CCs-DNPH-d3) were drawn in a 1:1 ratio, mixed thoroughly, and filtered through a 0.22 μm organic phase filter to obtain the third tobacco sample;

[0062] Step 5: The blank sample was prepared according to the above steps, except that the water infiltration step of the tobacco raw material was omitted;

[0063] Step 6: Perform UPLC-Orbitrap-HRMS analysis on the first tobacco sample, the second tobacco sample, the third tobacco sample, and the blank sample to obtain raw LC-MS data;

[0064] The conditions for UPLC-Orbitrap-HRMS analysis were:

[0065] UPLC-HRMS was performed using an instrument platform consisting of a Dionex U3000 UHPLC system and a Q–Exactive mass spectrometer in tandem;

[0066] UPLC conditions: Syncronis C18 column (2.1 mm × 100 mm, 1.7 μm); column temperature: 40°C; injection volume: 1 μL; mobile phase A: 0.1% formic acid in water; mobile phase B: acetonitrile; gradient elution program: 95% B at 0–1 min, 95%–60% B at 1–3 min, 60%–10% B at 3–10 min, 10% B at 10–18 min, 10% B at 18–19 min, 10%–95% B at 19–20 min, 95% B; flow rate: 0.2 mL / min;

[0067] Mass spectrometry conditions included a spray voltage of 3.7 kV, a sheath gas flow rate of 35 L / min, an auxiliary gas flow rate of 10 L / min, and a DL transfer tube temperature of 350°C. Data acquisition was performed in the negative ion mode of Full MS-DDA. The mass-to-charge ratio scan range for the primary mass spectrometer was set to 100–1200 m / z at a resolution of 70,000. The mass-to-charge ratio scan resolution for the secondary mass spectrometer was 35,000, and the high-energy collision-induced dissociation voltage was 30 eV.

[0068] Step 7: Perform peak detection, peak alignment, and peak grouping on the raw LC-MS data to obtain the mass spectral characteristics of the coexistence of target components and interfering components, including m / z values, retention times, and peak intensities;

[0069] Step 8: performing multiple filtering on the mass spectrum characteristic data of step 7, including: filtering the mass spectrum characteristic data based on statistical features, filtering the mass spectrum characteristic data based on mass defect, filtering the mass spectrum characteristic data based on paired chromatographic peaks marked by derivatization reactions, and filtering based on the mass spectrum characteristics of multiple product ions of the secondary mass spectrum, to obtain the final retained chromatographic peaks;

[0070] Among them: The steps of filtering mass spectrum feature data based on statistical features are:

[0071] Calculate the coefficient of variation CV and fold change FC, the calculation formula is CV = Int SQC / Int MQC ×100%; where Int SQC Represents the standard deviation of the peak intensity of a specific chromatographic peak in the first tobacco sample, Int MQC represents the average value of the peak intensity;

[0072] FC=Int MQC / Int MBK , where Int MQC Represents the average peak intensity of a specific chromatographic peak in the first tobacco sample, Int MBK represent the average values ​​of peak intensities of specific chromatographic peaks in blank samples;

[0073] If CV ≤ 30% and FC ≥ 1.5, retain the corresponding chromatographic peak;

[0074] Among them: The steps of filtering mass spectrum characteristic data based on mass defect are:

[0075] The calculation formula is: MD = |MZ–ceiling(MZ)|, where MZ represents the exact mass, i.e., the m / z value, of the precursor ion of the specific chromatographic peak detected in the first tobacco sample, and ceiling(MZ) is the ceiling number of the exact mass, i.e., the nominal mass;

[0076] If 0.02≤MD≤0.3 and m / z>209, retain the corresponding chromatographic peak.

[0077] The mass spectrum characteristic data filtering step of the paired chromatographic peaks marked by the derivatization reaction is as follows: a pair of chromatographic peaks detected in the third tobacco sample at the same retention time: if |MZ–MZ d3 |=3.01605, and |IntP1–IntP2| / max(P1,P2)<30%, and |RT1–RT2|<2, then retain the corresponding chromatographic peak; among them, MZ and MZ d3 where IntP1 represents the exact mass, i.e., the m / z value, of the precursor ions of the non-isotope-labeled and isotope-d3-labeled derivatization products, respectively; IntP1 and IntP2 represent the peak areas of the first and second chromatographic peaks, respectively; and RT1 and RT2 represent the retention times of the first and second chromatographic peaks, respectively.

[0078] Among them: the filtering method based on the secondary mass spectrometry multiple ion mass spectrometry characteristics is: if the secondary mass spectrum of the first tobacco sample can produce m / z 76.018, m / z 120.008, m / z 122.024, m / z 135.019 and m / z 181.012 fragment ions, retain the corresponding chromatographic peaks.

[0079] Step 9: Structural annotation or identification of the final retained chromatographic peaks using methods including standard matching, database retrieval, or fragmentation pattern analysis.

[0080] The above steps 7, 8 and 9 are automatically performed by the software package.

[0081] The present invention develops a data processing package (MSFiltering package) based on R language, which is used to perform peak detection, peak alignment and peak grouping processing on raw LC-MS data to obtain mass spectral characteristics of the coexistence of target components and interfering components. A multidimensional data filtering method based on statistical characteristics, mass defect, isotope labeling and multiple ion information of secondary mass spectrometry is used, and structural annotation or identification of the finally retained chromatographic peaks is performed through automated processing to achieve efficient recognition and identification of aldehydes and ketones in complex LC-MS data sets.

[0082] Figure 4 is a schematic diagram of the number of characteristic peaks of the tobacco sample in the embodiment, and the number of characteristic peaks after filtering the mass spectrum characteristic data based on statistical characteristics, the number of characteristic peaks after filtering the mass spectrum characteristic data based on mass loss, the number of characteristic peaks after filtering the mass spectrum characteristic data of paired chromatographic peaks labeled by derivatization reaction, and the number of characteristic peaks after filtering based on the multiple ion mass spectrum characteristics of secondary mass spectrometry.

[0083] As shown in Figure 4, the tobacco sample originally had 12,601 characteristic peaks. After statistical feature filtering (SCF), the number of retained peaks was reduced to 3,452 (a decrease of 72.6%). After mass defect filtering (MDF), the number of retained peaks was further reduced to 975 (a decrease of 92.3%). After derivatization-labeled paired peak filtering (PPF), the number of retained peaks was further reduced to 169 (a decrease of 98.6%). After secondary mass spectrometry multiple ion feature filtering (DFIF), the number of retained peaks was reduced to 93 (a decrease of 99.3%). Seventy peaks were retained by evaluating their EIC quality, molecular formula prediction, and primary and secondary mass spectrometry qualities. Structural identification of these retained peaks was performed by comparing retention times, MS, and MS / MS data based on comparison with standards and literature or database searches. For unknown compounds without standards or known mass spectrometric information, Frontier 7.0 (Thermo Fisher Scientific) was used to analyze mass spectrometric fragmentation patterns and annotate their chemical structures. The identification results are shown in Table 1, including molecular formula, MS / MS spectrum, and annotation information. Forty chromatographic peaks have structural annotations, and the chemical structures of another 20 peaks remain to be deduced. The annotated peaks include hydrazones (DNPH-derivatized products) of aldehydes or ketones such as formaldehyde, acetaldehyde, 2,3-butanedione, methylglyoxal, hydroxyacetone, furfural, 5-hydroxymethylfurfural, benzaldehyde, and salicylaldehyde.

[0084] The results show that there is a large amount of redundant information in the LC-MS raw data. The method of the present invention can effectively eliminate interfering information through multiple mass spectrometry filtering and quickly identify potential aldehyde and ketone chemical components from non-targeted data sets.

[0085] Comparative Example 1: The mass spectrometry characteristic data of the blank sample prepared in Example were subjected to the same multiple filtering, and the results are shown in Figure 1; as can be seen from Figure 1, the number of original characteristic peaks of the blank sample reached 7372; after filtering the mass spectrometry characteristic data based on statistical characteristics, the number of retained chromatographic peaks was reduced to 641; after filtering the mass spectrometry characteristic data based on mass loss, the number of retained chromatographic peaks was reduced to 51; after filtering the mass spectrometry characteristic data of paired chromatographic peaks labeled based on derivatization reaction, the number of retained chromatographic peaks was reduced to 0.

[0086] Comparative Example 2: The mass spectrometry characteristic data of the mixed standard sample are subjected to the same multiple filtering. The preparation steps of the mixed standard sample are the same as those in the embodiment. Standards of 24 aldehydes and 24 ketones are added to the mixed standard sample; the 24 aldehydes are: formaldehyde, acetaldehyde, acrolein, glyoxal, propanal, crotonaldehyde, malondialdehyde, butyraldehyde, valeraldehyde, furfural, glutaraldehyde, hexanal, benzaldehyde, 5-methylfurfural, heptaldehyde, phenylacetaldehyde, salicylaldehyde, 1-octanal, trans-cinnamaldehyde, 2,5-dimethylbenzaldehyde, p-methoxybenzaldehyde, 2,4-nonadienal, 2,4-condienal, condimental; the 24 ketones are: acetone, cyclopentanone, 2,3-butanedione, 3-methyl-2-cyclopentanedione, Pentenone, cyclohexanone, 2-methyltetrahydrofuran-3-one, 3-oxene-2-one, 4-heptanone, acetophenone, 2,3-heptanedione, isophorone, α-ionone, hydroxyacetone (acetol), 2-pentanone, acetoin, 2,3-pentanedione, methyl isobutyl ketone, 2,3-hexanedione, 2-heptanone, acetoxy-2-propanone, 6-methyl-3,5-oxene-2-one, 6-methyl-6-oxene-2-one, 4-methylacetophenone, 5-nonanone; the concentration of each is 0.1 mg / mL. The results of multiple filtering of the mass spectral characteristic data of the mixed standard sample are shown in Figure 2. As can be seen from Figure 2, the original number of characteristic peaks of the mixed standard sample reached 8021. After filtering the mass spectral characteristic data based on statistical features, the number of retained chromatographic peaks was reduced to 2440; after filtering the mass spectral characteristic data based on mass defect, the number of retained chromatographic peaks was reduced to 377; after filtering the mass spectral characteristic data based on paired chromatographic peaks labeled by derivatization reaction, the number of retained chromatographic peaks was reduced to 109; after filtering the mass spectral characteristic data based on paired chromatographic peaks labeled by derivatization reaction, the number of retained chromatographic peaks was reduced to 46. By observing the reconstructed TIC after filtering, many interfering mass spectral signals were eliminated, and the chromatographic peaks of the target were significantly exposed, which were not visible in the original TIC. This sharp reduction in the number of interfering mass spectral characteristics effectively narrowed the range of aldehyde and ketone components for identification. The recall rates obtained by filtering the mass spectral features of the mixed standard samples using different filtering methods are shown in Figure 3. As can be seen from Figure 3, the recall rates of the original data set, statistical feature filtering, and statistical feature combined with mass defect filtering are all 100%, and the recall rate of statistical feature combined with mass defect, paired chromatographic peaks and multiple product ion filtering is greater than 91%, indicating that the added standards can all be effectively identified, proving the effectiveness of the multiple filtering method.

[0087] The embodiments are merely preferred embodiments of the present invention and are not intended to limit the present invention. Those skilled in the art will readily appreciate that the present invention is susceptible to various modifications and variations. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention shall be included within the scope of protection of the present invention.

[0088] Table 1 Identification results of aldehydes and ketones based on UHPLC-Q-Orbitrap-MS / MS

Claims

1. A non-targeted mass spectrometry identification method for tobacco carbonyl components, characterized in that: The steps include: (1) soaking a tobacco sample with water, adding a certain amount of acetonitrile for oscillation extraction, and obtaining a tobacco sample extract; (2) taking a certain amount of tobacco sample extract, adding a certain amount of 2,4-dinitrophenylhydrazine (DNPH) thereto for derivatization, and preparing a first tobacco sample (CCs-DNPH); (3) using DNPH-d3 instead of DNPH, repeating step (2) to prepare a second isotope-labeled tobacco sample (CCs-DNPH-d3); (4) taking the first tobacco sample and the second tobacco sample in a ratio of 1:1, mixing them thoroughly, and filtering them through a filter membrane to obtain a third tobacco sample; (5) Prepare a blank sample according to steps (1) and (2); (6) performing UPLC-Orbitrap-HRMS analysis on the first tobacco sample, the second tobacco sample, the third tobacco sample and the blank sample respectively to obtain raw LC-MS data; (7) performing peak detection, peak alignment, and peak grouping on the raw LC-MS data to obtain several raw mass spectrometric features of the coexistence of target components and interfering components, including m / z values, retention times, and peak intensities; (8) performing multiple filtering on the original mass spectrum characteristic data obtained in (7); (9) Structural annotation or identification of the chromatographic peaks finally retained can yield tobacco carbonyl components.

2. The method according to claim 1, characterized in that The conditions for UPLC-Orbitrap-HRMS analysis were: UPLC-HRMS was performed using an instrument platform consisting of a Dionex U3000 UHPLC system and a Q–Exactive mass spectrometer in tandem; UPLC conditions: Syncronis C18 column (2.1 mm × 100 mm, 1.7 μm); column temperature: 40 °C; injection volume: 1 μL; mobile phase A: 0.1% formic acid in water; mobile phase B: acetonitrile; gradient elution program: 0-1 min 95% B, 1-3 min 95%-60% B, 3-10 min 60%-10% B, 10-18 min 10% B, 18-19 min 10%-95% B, 19-20 min 95% B; Flow rate: 0.2 mL / min; Mass spectrometry conditions: spray voltage 3.7 kV, sheath gas flow rate 35 L / min, auxiliary gas flow rate 10 L / min, DL transfer tube temperature 350 ° C. Data acquisition was carried out in the negative ion mode of Full MS-DDA, and the mass-to-charge ratio scanning range of the primary mass spectrometer was set to: 100-1200 m / z, with a resolution of 70000; the mass-to-charge ratio scanning resolution of the secondary mass spectrometer was 35000, and the high-energy collision-induced dissociation voltage was 30 eV.

3. The method according to claim 1, characterized in that Multiple filtering of mass spectral feature data includes: mass spectral feature data filtering based on statistical features, mass spectral feature data filtering based on mass loss, mass spectral feature data filtering of paired chromatographic peaks based on derivatization reaction labels, and mass spectral feature filtering based on secondary mass spectrometry multiple ion to obtain the final retained chromatographic peaks.

4. The method according to claim 3, characterized in that The steps of filtering mass spectrum feature data based on statistical features are: Calculate the coefficient of variation CV and fold change FC, the calculation formula is CV = Int SQC / Int MQC ×100%; where Int SQC Represents the standard deviation of the peak intensity of a specific chromatographic peak in the first tobacco sample, Int MQC represents the average value of the peak intensity; FC=Int MQC / Int MBK , where Int MQC Represents the average peak intensity of a specific chromatographic peak in the first tobacco sample, Int MBK They represent the average values ​​of peak intensities of specific chromatographic peaks in blank samples, respectively; If CV≤30% and FC≥1.5, retain the corresponding chromatographic peak.

5. The method according to claim 3, characterized in that: The steps for filtering mass spectral feature data based on mass defect are: The calculation formula is: MD = |MZ–ceiling(MZ)|, where MZ represents the exact mass, i.e., the m / z value, of the precursor ion of the specific chromatographic peak detected in the first tobacco sample, and ceiling(MZ) is the ceiling number of the exact mass, i.e., the nominal mass; If 0.02≤MD≤0.3, m / z>209, retain the corresponding chromatographic peak.

6. The method according to claim 3, characterized in that The steps of filtering the mass spectrum characteristic data of the paired chromatographic peaks labeled by the derivatization reaction are as follows: a pair of chromatographic peaks detected in the third tobacco sample at the same retention time: if |MZ–MZ d3 |=3.0186, and |IntP1–IntP2| / max(P1,P2)<30%, and |RT1–RT2|<2, then the corresponding chromatographic peak is retained; among them, MZ and MZ d3 Represent the accurate mass, i.e., m / z value, of the precursor ions of the non-isotope labeled and isotope d3 labeled derivatization products, respectively; IntP1 and IntP2 represent the peak areas of the first and second chromatographic peaks, respectively; RT1 and RT2 represent the retention times of the first and second chromatographic peaks, respectively.

7. The method according to claim 3, characterized in that The filtering method based on the secondary mass spectrometry multiple ion mass spectrometry characteristics is: if the secondary mass spectrum of the first tobacco sample can produce m / z 76.018, m / z 120.008, m / z 122.024, m / z 135.019 and m / z 181.012 fragment ions, the corresponding chromatographic peaks are retained.

8. The method according to claim 3, characterized in that The methods used for structural annotation or identification of the finally retained chromatographic peaks include: standard sample matching, database search, or fragmentation pattern analysis.

9. The method according to claim 1, characterized in that: Step (7) performs peak detection, peak alignment and peak grouping processing, step (8) performs multiple filtering on mass spectrum feature data and step (9) performs structural annotation or identification on the chromatographic peaks finally retained, which are automatically performed by the software package.

Images