Metabolin qualitative annotation quality evaluation method based on high-resolution mass spectrum simulation

Abstract

The invention provides a metabolite qualitative annotation quality evaluation method based on high-resolution mass spectrum simulation. The metabolite qualitative annotation quality evaluation method comprises the following steps: collecting human serum sample mgf data measured by different instrument platforms; determining the distribution characteristic and the intensity mode of the mass-to-charge ratio; constructing a high-resolution mass spectrum data model based on mass spectrum characteristic quantity, adjacent m / z difference and intensity distribution; constructing a high-resolution mass spectrum simulation database of hundreds of thousands of metabolites with similarity and difference; constructing a strong interference strategy and a weak interference strategy, and simulating mass spectrum data change possibly encountered in an actual experiment; searching and evaluating an undetermined property spectrum under the interference model by adopting eight similarity evaluation methods; factor level analysis and range analysis are adopted to deeply discuss the internal relation between the scoring algorithm and the mass spectrum characteristics.

Description

Technical Field

[0001] The present invention relates to the technical field of analytical chemistry, and in particular to a metabolite qualitative annotation quality evaluation method based on high-resolution mass spectrometry simulation. Background Art

[0002] Metabolomics studies substances with a molecular weight of less than 1,500 Daltons in cells, tissues or organisms. By systematically characterizing and quantitatively measuring these small molecules, it describes the dynamic changes in their physiological or pathological processes, and then studies the phenomena and intrinsic nature of life processes, explains the mechanisms, and solves complex and diverse research and application challenges.

[0003] Qualitative annotation of metabolites is at the core of the entire metabolomics research process. Its goal is to accurately identify and confirm the chemical structure and molecular characteristics of metabolites detected in biological samples. This step is crucial for subsequent metabolic pathway analysis, mechanism analysis, and verification analysis. With the rapid development of instrument hardware, the qualitative identification technology of metabolites has made great progress in the past decade. High-resolution mass spectrometry has become the tool of choice for non-targeted metabolomics research with its excellent resolution, selectivity, and detection sensitivity. In particular, the combination of ultra-high performance liquid chromatography and high-resolution mass spectrometry (UPLC-HRMS) provides a powerful platform for the detection of metabolites. In LC-HRMS detection, qualitative information such as chromatographic retention time (RT), full scan mass-to-charge ratio (m / z), and secondary mass spectrometry (MS / MS) are key elements of annotation analysis.

[0004] However, to achieve the gold standard for qualitative identification of metabolites, there are still a series of difficulties and challenges. Due to the complexity of mass spectrometry data, the diversity of experimental conditions and the lack of standards, it is challenging to build a universal database and accurately annotate metabolites. In existing public databases such as HMDB, Metlin and PubChem, more than 90% of metabolites do not have corresponding standard MS / MS mass spectrometry data, which limits the breadth and depth of qualitative annotation. In addition, the fragmentation spectrum of the same compound is affected by multiple factors such as instrument type, ion source, fragmentation mode, experimental parameters and matrix effect, resulting in a high degree of conditional dependence of MS / MS data obtained under different experimental conditions. The accuracy of database-based search results is not only related to the quality of the mass spectra included in the spectral library, but also to the specific scoring formula. Different similarity search methods may lead to significant differences in qualitative results. Therefore, the existing technology is insufficient in the accuracy and reliability of qualitative identification of metabolites. Summary of the invention

[0005] According to the technical problems mentioned in the above background technology, a metabolite qualitative annotation quality evaluation method based on high-resolution mass spectrometry simulation is provided.

[0006] The technical means adopted by the present invention are as follows:

[0007] A metabolite qualitative annotation quality evaluation method based on high-resolution mass spectrometry simulation comprises the following steps:

[0008] S1: Construction of simulated high-resolution mass spectrometry database;

[0009] S2: Construct strong and weak interference models to simulate mass spectrometry data with different degrees of variation encountered in actual experiments;

[0010] S3: Eight similarity evaluation methods were used to compare the search results of the mass spectra to be identified in the simulated database under the interference model to identify the potential for accurately finding metabolic annotation results under different interference conditions.

[0011] Furthermore, the construction of the simulated high-resolution mass spectrometry database comprises the following steps:

[0012] S11: Collect human serum sample measurement data, respectively at AB SCIEX 5600+, Agilent6546HPLC / Q-TOF and Thermo Scientific TM Q Exactive TM Under the three instruments of HF UPLC / Q-Orbitrap, the raw secondary mass spectrometry data of the compounds in mgf format were obtained to determine the distribution characteristics of the mass-to-charge ratio m / z and the corresponding intensity pattern;

[0013] S12: Based on the three key factors of mass spectral feature number, adjacent m / z differences (i.e., the proportion of peaks with m / z value difference less than 0.05Da in the total number of peaks), and feature intensity distribution, the {metabolite1:[[m / z1,intensity1],[m / z2,intensity2]...],metabolite2:...} data structure was constructed using Python programming language to comprehensively characterize metabolite structures of different complexity.

[0014] S13: Using the orthogonal experimental design method, three levels were set for each factor to generate nine parameter combinations, and 1000 corresponding data sets were randomly generated through the simulation algorithm to produce a sub-database containing 9000 mass spectrometry data;

[0015] S14: Structural similarity was introduced based on the mass spectrometry sub-database. For each of the 9000 mass spectrometry data generated, a 0.05- to 0.5-fold m / z drift and a -0.5 to 0.5-fold change in relative intensity were randomly introduced. The m / z drift window was set between -1Da and 1Da, thus generating 12 variants.

[0016] S15: Build a virtual high-resolution mass spectrometry database of 100,000 levels.

[0017] Furthermore, the construction of strong and weak interference models to simulate mass spectrometry data of different degrees of variation encountered in actual experiments includes the following steps:

[0018] S21: A strong interference strategy was adopted to simulate mass spectrum changes, and a mass-to-charge ratio increase or decrease of 0.1 to 0.2 units was randomly selected to simulate the mass-to-charge ratio changes that occurred under extreme experimental conditions;

[0019] S22: Introduce a weak interference simulation strategy, randomly select mass-to-charge ratios of 0.05 to 0.1 times below 30% of the base peak intensity for deletion, increase mass-to-charge ratios of 0.05 to 0.1 times, and use the Poisson distribution model to assign intensity values ​​to them to simulate background noise signal interference.

[0020] Furthermore, the eight similarity evaluation methods are used to compare the search results of the mass spectra to be identified in the simulation database under the interference model, and to identify their potential to accurately find metabolic annotation results under different interference conditions, including the following steps:

[0021] S31: 900 mass spectrometry data were randomly selected from the sub-database, strong and weak interference strategies were introduced to these mass spectra, mass spectra changes under extreme experimental conditions were simulated, and a test set of unidentified substances was generated;

[0022] S32: Apply 8 similarity evaluation methods to each MS / MS spectrum in the test set; the results of the 8 similarity evaluation methods are all in the interval [0,1], where 1 indicates a perfect positive correlation and 0 indicates no correlation;

[0023] S33: Evaluate the performance of the similarity method. The higher the accuracy of the mass spectrometry similarity search results, the better the performance. Use factor level analysis and range analysis to further explore the relationship between the scoring algorithm and mass spectrometry characteristics.

[0024] Furthermore, the eight similarity evaluation methods include: entropy similarity (ES), dot product (DP), weighted dot product (WDP), MS for ID (MS4ID), Pearson correlation coefficient (PCC), Bhattacharya similarity (BS), harmonic mean (HM), and Euclidean distance (ED).

[0025] Furthermore, the formulas for calculating similarity by the eight similarity evaluation methods are as follows:

[0026]

[0027] Where M represents the mass-to-charge ratio; I represents the intensity; and q represents that the test spectrum is a collection of peaks: (M q,1 ,I q,1 ),...,(M q,i ,I q,i ); r represents the library spectrum, defined as (M r,1 ,I r,1 ),...,(M r,i ,I r,i );N q 、N r Indicates the number of fragment ions of spectra q and r; N m represents the total number of matched fragment ions; represent the average intensity of the test spectrum and library spectrum, respectively;

[0028] Compared with the prior art, the present invention has the following advantages:

[0029] The present invention effectively improves the accuracy of qualitative annotation. By simulating the construction of a database and evaluating the similarity search algorithm, the present invention provides a new perspective and tool for solving the complexity problem of mass spectrometry data in HRMS analysis, thereby improving the accuracy of metabolite qualitative annotation. By introducing strong and weak interference simulation strategies, the present invention evaluates and optimizes the robustness of the metabolite qualitative annotation algorithm, so that it can better adapt to the mass spectrometry data changes that may be encountered in actual experiments.

[0030] The present invention not only provides a new strategy and method for evaluating the results of qualitative annotation of metabolites, but also has important scientific significance and application value for promoting the research of discovery metabolomics. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0032] Figure 1 Analyze the schematic diagram of the present invention.

[0033] Figure 2 The representative MS / MS spectra selected from the sub-database of the present invention include: (AC) 5-10 mass-to-charge ratios; (DF) 10-50 mass-to-charge ratios; (GI) 50-100 mass-to-charge ratios.

[0034] Figure 3 The comparison analysis diagrams of the standard mass spectrum and the actual sample mass spectrum in the database of the present invention. (A) is the mass spectrum comparison under strong interference conditions; (B) is the mass spectrum comparison under weak interference conditions;

[0035] Figure 4 The performance analysis diagram of the similarity algorithm of the present invention under strong interference and weak interference conditions. (A) is a search accuracy-ranking position diagram of eight similarity algorithms in the strong interference test. Each point represents the percentage of correct retrieval found in a certain number of best (top) retrievals; (B) is a box plot of the correct retrieval score of the top ranked similarity method in the strong interference test; (C) is a characteristic representation of the superiority of the WDPMS4IdHM algorithm compared to the ES algorithm in the strong interference test, and the value on the circle represents the intensity; (D) is a similarity algorithm search accuracy-ranking position diagram in the weak interference test; (E) is a box plot of the correct retrieval score of the top ranked similarity method in the weak interference test; (F) is a characteristic representation of the superiority of the WDPMS4IdHM algorithm compared to the ES algorithm in the weak interference test.

[0036] Figure 5 The comparison diagram of the impact of three factors on the performance of eight similarity algorithms in a strong interference environment of the present invention. (A) is a three-factor three-level statistical analysis; (B) is a range statistical analysis.

[0037] Figure 6 The comparison diagram of the impact of three factors on the performance of eight similarity algorithms in the weak interference environment of the present invention. (A) is a three-factor three-level statistical analysis; (B) is a range statistical analysis. DETAILED DESCRIPTION

[0038] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0039] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0040] like Figure 1 FIG. 1 is an analysis principle diagram of a metabolite qualitative annotation quality evaluation method based on high-resolution mass spectrometry simulation of the present invention, comprising the following steps:

[0041] S1: Construction of simulated high-resolution mass spectrometry database;

[0042] S2: Construct strong and weak interference models to simulate mass spectrometry data of different degrees of variation that may be encountered in actual experiments;

[0043] S3: Eight similarity methods were used to compare the search results of the mass spectra to be identified in the simulated database under the interference model to identify their potential to accurately find metabolic annotation results under different interference conditions.

[0044] Further: the construction of a simulated high-resolution mass spectrometry database comprises:

[0045] S11: Collect human serum sample measurement data, respectively at AB SCIEX 5600+, Agilent6546 HPLC / Q-TOF and Thermo Scientific TM Q Exactive TM The raw secondary mass spectrometry data of the compounds in mgf format were obtained under HF UPLC / Q-Orbitrap instruments to determine the distribution characteristics of the mass-to-charge ratio (m / z) and the corresponding intensity pattern.

[0046] S12: Based on the three key factors of mass spectral feature number, adjacent m / z variability (i.e., the proportion of peaks with m / z value difference less than 0.05Da in the total number of peaks), and feature intensity distribution, an accurate high-resolution mass spectrometry data simulation algorithm was developed using the Python programming language to construct the {metabolite1: [[m / z1, intensity1], [m / z2, intensity2]...], metabolite2: ...} data structure to comprehensively characterize metabolite structures of different complexities;

[0047] S13: An orthogonal experimental design method was used to set three levels for each factor, generate nine parameter combinations, and randomly generate 1,000 corresponding data sets through a simulation algorithm, ultimately generating a sub-database containing 9,000 mass spectrometry data;

[0048] S14: Structural similarity was further introduced based on the mass spectrometry sub-database. For each of the 9,000 mass spectrometry data, a 0.05- to 0.5-fold m / z drift and a -0.5 to 0.5-fold change in relative intensity were randomly introduced. The m / z drift window was set between -1Da and 1Da, thus generating 12 variants.

[0049] S15: Integrating S13-S14, we finally generated 117,000 substances and built a virtual high-resolution mass spectrometry database of 100,000 levels.

[0050] Further: the construction of strong and weak interference models to simulate mass spectrometry data of different degrees of variation that may be encountered in actual experiments includes:

[0051] S21: A strong interference strategy was adopted to simulate the mass spectrum changes that may be caused by factors such as ion source, instrument type, collision energy, and data acquisition mode. The mass-to-charge ratio increase or decrease of 0.1 to 0.2 units was randomly selected to simulate the mass-to-charge ratio changes that may occur under extreme experimental conditions.

[0052] S22: A weak interference simulation strategy was proposed, in which the mass-to-charge ratios (m / z) of 0.05 to 0.1 times below 30% of the base peak intensity were randomly selected for deletion, the mass-to-charge ratios of 0.05 to 0.1 times were added, and the Poisson distribution model was used to assign intensity values ​​to them to simulate background noise signal interference.

[0053] Further: the search results of using eight similarity methods to compare the mass spectra to be determined in the simulation database under the interference model include:

[0054] S31: 900 mass spectrometry data were randomly selected from the sub-database, strong and weak interference strategies were introduced to these mass spectra, mass spectra changes under extreme experimental conditions were simulated, and a test set of unidentified substances was generated;

[0055] S32: Eight similarity evaluation methods are applied to each MS / MS spectrum in the test set, including entropysimilarity(ES), dot product(DP), weighted dot product(WDP), MS for ID(MS4ID), Pearson correlation coefficient(PCC), Bhattacharya similarity(BS), harmonicmean(HM), Euclidean distance(ED) to calculate similarity(1)-(8):

[0056]

[0057]

[0058] Where M represents the mass-to-charge ratio; I represents the intensity; and q represents that the test spectrum is a collection of peaks: (M q,1 ,I q,1 ),...,(M q,i ,I q,i ); r represents the library spectrum, defined as (M r,1 ,I r,1 ),...,(M r,i ,I r,i );N q 、N r Indicates the number of fragment ions of spectra q and r; N m represents the total number of matched fragment ions; represent the average intensity of the test spectrum and library spectrum, respectively;

[0059] The results of the above eight similarity algorithms are all in the interval [0,1], where 1 indicates a perfect positive correlation and 0 indicates no correlation.

[0060] S33: Performance evaluation of similarity methods. The higher the accuracy of mass spectrometry similarity search results, the better the performance. Factor level analysis and range analysis are used to further explore the intrinsic relationship between the scoring algorithm and mass spectrometry characteristics.

[0061] Example 1

[0062] In this embodiment, respectively at AB SCIEX 5600+, Agilent 6546HPLC / Q-TOF and Thermo Scientific TM Q Exactive TMThe raw secondary mass spectrometry data of human serum in mgf format were obtained under HF UPLC / Q-Orbitrap instruments to determine the mass-to-charge ratio (m / z)

[0063] The orthogonal experimental design method was further used to construct a simulated high-resolution mass spectrometry database based on the three factors of feature number, the proportion of peaks with m / z value difference less than 0.05Da in the total number of peaks, and intensity distribution. The specific information is as follows:

[0064]

[0065] The above three-factor three-level orthogonal experiment finally produced a sub-database containing 9000 mass spectrometry data, such as Figure 2 shown.

[0066] Figure 2 These are representative MS / MS spectra selected from the sub-database, showing different features, with the number of mass-to-charge ratios ranging from: (AC) 5-10 mass-to-charge ratios; (DF) 10-50 mass-to-charge ratios; (GI) 50-100 mass-to-charge ratios.

[0067] The results show that the constructed database is representative and comprehensive, and each data set represents a different virtual substance. Then, 100 mass spectra were randomly selected for each parameter combination, and a total of 900 mass spectra were selected. Figure 3 As shown, the mass-to-charge ratio of 0.1 to 0.2 units was randomly selected for these mass spectra, and the strong interference strategy was used to simulate the mass spectrum changes under extreme experimental conditions; the weak interference simulation strategy was to randomly select 0.05 to 0.1 times the mass-to-charge ratio (m / z) below 30% of the base peak intensity for deletion, increase the mass-to-charge ratio by 0.05 to 0.1 times, and use the Poisson distribution model to assign intensity values ​​to them to simulate background noise signal interference. The results show that the difference between the actual undetermined information and the database data is simulated.

[0068] Example 2

[0069] In this embodiment, first, the strong and weak interference strategies constitute a test set of substances to be identified, and 8 similarity evaluation methods are applied to each MS / MS spectrum in the test set, including entropy similarity (ES), dot product (DP), weighted dot product (WDP), MS for ID (MS4ID), Pearson correlation coefficient (PCC), Bhattacharya similarity (BS), harmonic mean (HM), Euclidean distance (ED), and searched in the simulation database. The similarity score is between 0 and 1, and the search results are output to an Excel table. The results are analyzed as follows Figure 4 shown.

[0070] Figure 4 It is a performance analysis chart of similarity algorithms under strong interference and weak interference conditions. (A) Search accuracy-ranking position chart of eight similarity algorithms in strong interference test. Each point represents the percentage of correct retrieval found in a certain number of best (top) retrievals; (B) Box plot of correct retrieval scores of the top eight similarity methods in strong interference test; (C) Characterization of the superiority characteristics of WDP, MS4Id, and HM algorithms compared to ES algorithm in strong interference test, the values ​​on the circles represent the strength; (D) Search accuracy-ranking position chart of similarity algorithms in weak interference test; (E) Box plot of correct retrieval scores of the top similarity methods in weak interference test; (F) Characterization of the superiority characteristics of WDP, MS4Id, and HM algorithms compared to ES algorithm in weak interference test.

[0071] The performance of eight mass spectrometry similarity evaluation methods was further analyzed by factor level analysis and range analysis to reveal the performance differences of different algorithms in processing mass spectrometry data and explore the factors that affect these performances, such as Figure 5 , Figure 6 shown.

[0072] Figure 5 Comparison of the impact of three factors on the performance of eight similarity algorithms under strong interference environment (A) Three-factor three-level statistical analysis: (B) Range statistical analysis. Figure 6 Comparison of the impact of three factors on the performance of eight similarity algorithms in a weak interference environment (A) Three-factor three-level statistical analysis: (B) Extreme difference statistical analysis.

[0073] The results show that the performance of the similarity evaluation method is related to the specific experimental conditions. When the actual operating experimental conditions are slightly different from those in the database, the algorithm performance is generally good. The ES algorithm shows strong anti-interference ability, especially when processing high-dimensional data, its annotation recognition ability is significantly better than other algorithms, and it is recommended for priority use. In addition, the BS and HM algorithms also have strong stability, but when there are a large number of close mass signals, the algorithm performance deteriorates. The mass-to-charge ratio is the main factor affecting the performance of mass spectrometry similarity evaluation methods, and different algorithms have different sensitivities to the m / z characteristic distribution and intensity changes of mass spectrometry data. The MS4ID algorithm has poor overall performance, but it can effectively supplement the shortcomings of the ES algorithm when processing mass spectra with less than 10 product ion fragments.

[0074] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages or disadvantages of the embodiments.

[0075] In the above embodiments of the present invention, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0076] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only schematic. For example, the division of the units can be a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0077] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple units. Some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0078] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0079] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions to enable a computer device (which can be a personal computer, server or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, disk or optical disk and other media that can store program codes.

[0080] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A metabolite qualitative annotation quality evaluation method based on high-resolution mass spectrometry simulation, characterized in that: The following steps are involved: S1: Construction of simulated high-resolution mass spectrometry database; S2: Construct strong and weak interference models to simulate mass spectrometry data with different degrees of variation encountered in actual experiments; S3: Eight similarity evaluation methods were used to compare the search results of the mass spectra to be identified in the simulated database under the interference model to identify the potential for accurately finding metabolic annotation results under different interference conditions.

2. A metabolite qualitative annotation quality evaluation method based on high-resolution mass spectrometry simulation according to claim 1, characterized in that: The construction of the simulated high-resolution mass spectrometry database comprises the following steps: S11: Collect human serum sample measurement data, respectively at AB SCIEX 5600+, Agilent 6546HPLC / Q-TOF and Thermo Scientific TM Q Exactive TM Under the three instruments of HF UPLC / Q-Orbitrap, the raw secondary mass spectrometry data of the compounds in mgf format were obtained to determine the distribution characteristics of the mass-to-charge ratio m / z and the corresponding intensity pattern; S12: Based on the three key factors of mass spectral feature number, adjacent m / z differences (i.e., the proportion of peaks with m / z value difference less than 0.05Da in the total number of peaks), and feature intensity distribution, the {metabolite1:[[m / z1,intensity1],[m / z2,intensity2]...],metabolite2:...} data structure was constructed using Python programming language to comprehensively characterize metabolite structures of different complexity. S13: Using the orthogonal experimental design method, three levels were set for each factor to generate nine parameter combinations, and 1000 corresponding data sets were randomly generated through the simulation algorithm to produce a sub-database containing 9000 mass spectrometry data; S14: Structural similarity was introduced based on the mass spectrometry sub-database. For each of the 9000 mass spectrometry data generated, a 0.05- to 0.5-fold m / z drift and a -0.5 to 0.5-fold change in relative intensity were randomly introduced. The m / z drift window was set between -1Da and 1Da, thus generating 12 variants. S15: Build a virtual high-resolution mass spectrometry database of 100,000 levels.

3. The method for qualitative annotation quality evaluation of metabolites based on high-resolution mass spectrometry simulation according to claim 1, characterized in that: The construction of strong and weak interference models to simulate mass spectrometry data of different degrees of variation encountered in actual experiments includes the following steps: S21: A strong interference strategy was adopted to simulate mass spectrum changes, and a mass-to-charge ratio increase or decrease of 0.1 to 0.2 units was randomly selected to simulate the mass-to-charge ratio changes that occurred under extreme experimental conditions; S22: Introduce a weak interference simulation strategy, randomly select mass-to-charge ratios of 0.05 to 0.1 times below 30% of the base peak intensity for deletion, increase mass-to-charge ratios of 0.05 to 0.1 times, and use the Poisson distribution model to assign intensity values ​​to them to simulate background noise signal interference.

4. The method for qualitative annotation quality evaluation of metabolites based on high-resolution mass spectrometry simulation according to claim 1, characterized in that: The eight similarity evaluation methods are used to compare the search results of the mass spectra to be identified in the simulation database under the interference model, and to identify their potential to accurately find metabolic annotation results under different interference conditions, including the following steps: S31: 900 mass spectrometry data were randomly selected from the sub-database, strong and weak interference strategies were introduced to these mass spectra, mass spectra changes under extreme experimental conditions were simulated, and a test set of unidentified substances was generated; S32: 8 similarity evaluation methods are applied to each MS / MS spectrum in the test set; the results of the 8 similarity evaluation methods are all in the interval [0,1], where 1 indicates a perfect positive correlation and 0 indicates no correlation; S33: Evaluate the performance of the similarity method. The higher the accuracy of the mass spectrometry similarity search results, the better the performance. Use factor level analysis and range analysis to further explore the relationship between the scoring algorithm and mass spectrometry characteristics.

5. A metabolite qualitative annotation quality assessment method based on high-resolution mass spectrometry simulation according to claim 1 or 4, characterized in that: The eight similarity evaluation methods include: entropy similarity (ES), dot product (DP), weighted dot product (WDP), MS for ID (MS4ID), Pearson correlation coefficient (PCC), Bhattacharya similarity (BS), harmonic mean (HM), and Euclidean distance (ED).

6. A metabolite qualitative annotation quality assessment method based on high-resolution mass spectrometry simulation according to claim 1 or 4, characterized in that: The formulas for calculating similarity of the eight similarity evaluation methods are as follows: Where M represents mass-to-charge ratio; I represents intensity; q represents that the test spectrum is a collection of peaks: (M q,1 ,I q,1 ),...,(M q,i ,I q,i ); r represents the library spectrum, defined as (M r,1 ,I r,1 ),...,(M r,i ,I r,i );N q 、N r Indicates the number of fragment ions of spectra q and r; N m represents the total number of matched fragment ions; represent the average intensity of the test spectrum and library spectrum, respectively;

Images