Intelligent analysis system and method for identifying dimer isomers in beta-lactam antibiotics

Abstract

The invention discloses an intelligent analysis system for identifying dimer isomers in beta-lactam antibiotics, which is characterized in that a potential characteristic fragment data collection subsystem is used for acquiring each dimer isomer and potential characteristic fragment data of the dimer isomer, which are theoretically calculated, and the potential characteristic fragment data comprise the molecular weight of potential characteristic fragments in a positive ion mode; screening the potential feature fragment data of each dimer isomer by a unique feature fragment data screening subsystem to obtain corresponding unique feature fragment data; the mass spectrum image recognition subsystem performs target mass spectrum image selection under the target molecular weight on the detected mass spectrum images, and extracts mass spectrum image data including mass-to-charge ratios corresponding to fragment ion peaks; and comparing each unique characteristic fragment data with each mass spectrum image identification data by a comparison subsystem, and obtaining a matching result of the molecular weight and the mass-to-charge ratio according to a preset fragment matching strategy. The biopolymer isomer impurity screening device can replace detection personnel to effectively screen the biopolymer isomer impurities.

Description

Technical Field

[0001] The invention relates to the field of drug analysis and quality control, and in particular to an intelligent analysis system and method for dimer isomers in beta-lactam antibiotics. Background Art

[0002] β-lactam antibiotics (β-lactams) refer to a large class of antibiotics with a β-lactam ring in their chemical structure. They are widely used in clinical practice to treat sepsis, pneumonia, lung abscesses, secondary infections caused by chronic respiratory diseases, refractory cystitis, pyelonephritis, peritonitis, gynecological adnexitis, etc. caused by sensitive bacteria. Although this product has a broad antibacterial spectrum and strong antibacterial activity, it has extremely poor stability and is easily degraded when exposed to moisture and heat, producing highly allergenic polymer impurities, which seriously threaten the life safety of patients.

[0003] At present, most of the analyses of allergenic polymer impurities in β-lactam antibiotics are based on instrumental quantitative analysis. The structure of polymer impurities has not been clearly identified, and the detected polymer impurities are limited. Multiple studies have found that isomeric impurities in polymer impurities are closely related to their toxic effects; European Pharmacopoeia EP 8.0 and United States Pharmacopoeia USP 37 / NF32 include several dimer isomers, reflecting the importance of isomer identification. However, the structure of isomeric impurities is relatively close to that of the main components of the drug, which is a difficult point in impurity identification. Therefore, it is necessary to develop new pathways to study the potential polymerization sites of polymer impurities and identify the structural characteristics of isomers in order to control impurities in β-lactam antibiotics. Summary of the invention

[0004] The purpose of the present invention is to provide an intelligent analysis system and method for identifying dimer isomers in β-lactam antibiotics, so as to solve the problem of difficulty in impurity identification due to the close structure of dimer isomer impurities and the main component of the drug.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] An intelligent analysis system for identifying dimer isomers in β-lactam antibiotics, comprising:

[0007] The potential characteristic fragment data collection subsystem is used to intelligently read dimer isomers and their potential characteristic fragments using theoretical cleavage fragments of dimer impurities, and obtain theoretically calculated potential characteristic fragment data of each dimer isomer, including the molecular weight of the potential characteristic fragment in positive ion mode;

[0008] The unique characteristic fragment data screening subsystem is used to screen the potential characteristic fragment data of each dimer isomer to obtain the corresponding unique characteristic fragment data, including the molecular weight of the unique characteristic fragment in the positive ion mode;

[0009] The mass spectrum image recognition subsystem is used to select the target mass spectrum image under the target molecular weight from the mass spectrum image obtained by the actual dimer impurity detection, and extract its mass spectrum image data, including the mass-to-charge ratio corresponding to the fragment ion peak;

[0010] The comparison subsystem is used to compare the theoretically calculated unique characteristic fragment data with the actual mass spectrum image recognition data of the dimer impurity, and obtain the matching result for the molecular weight and mass-to-charge ratio according to the preset fragment matching strategy.

[0011] Preferably, the potential feature segment data collection subsystem includes:

[0012] A first data reading module is used to read theoretical cleavage fragments of dimer impurities;

[0013] The feature collection module is used to list the polymerization modes based on each theoretical cleavage fragment and its different potential polymerization sites, obtain several different dimer isomers and collect relevant feature data;

[0014] A calculation module, for processing relevant characteristic data for each potential characteristic fragment set of each dimer isomer, wherein the potential characteristic fragment is a fragment obtained by cutting the dimer isomer according to the cutting rule of the carbapenem compound while keeping the amide or ester bond in the dimer connecting the two molecules intact;

[0015] The first data output module is used to output each dimer isomer and its potential characteristic fragment data, each potential characteristic fragment data includes the molecular weight of each potential characteristic fragment in the positive ion mode.

[0016] Preferably, the unique feature fragment data screening subsystem includes:

[0017] A second data reading module is used to read potential characteristic fragment data of each dimer isomer;

[0018] The feature screening module is used to eliminate the potential feature fragments common to other dimer isomers from the potential feature fragment data, and obtain the unique feature fragment data of each dimer isomer;

[0019] The second data output module is used to output the unique characteristic fragment data of each dimer isomer, including the molecular weight of the unique characteristic fragment in the positive ion mode.

[0020] Preferably, the mass spectrometry image recognition subsystem comprises:

[0021] A third data reading module is used to read the mass spectrum image actually detected;

[0022] The image recognition and conversion module is used to determine the RT range of the dimer impurity in the total ion current map based on the target molecular weight, obtain the mass spectrum image within the RT range as the target mass spectrum image, and extract mass spectrum image recognition data for any target mass spectrum image, including extracting the mass-to-charge ratio and mass spectrum peak intensity in the positive ion mode;

[0023] The third data output module is used to output each mass spectrum image recognition data within the target RT range.

[0024] Preferably, in the image recognition module, the number of target mass spectrometry images is no less than 3.

[0025] Preferably, the comparison subsystem comprises:

[0026] A fourth data reading module, used to read each unique characteristic fragment data and each mass spectrum image recognition data;

[0027] The accuracy control module is used to set the fragment matching strategy and its matching accuracy, wherein the fragment matching strategy includes a first matching strategy and a second matching strategy; in the first matching strategy, when the absolute difference between the molecular weight and the mass-to-charge ratio does not exceed a preset value, the corresponding unique characteristic fragment data is determined to match the mass spectrum image recognition data; in the second matching strategy, when the same unique characteristic fragment data matches no less than two mass spectrum image recognition data in the same RT interval, the corresponding dimer isomer is determined to be consistent with the dimer impurity in the RT interval.

[0028] A comparison module, for matching the molecular weight of the unique characteristic fragment data and the mass-to-charge ratio of the mass spectrum image recognition data according to a preset fragment matching strategy;

[0029] The fourth output module is used to output the matching result.

[0030] Preferably, the comparison subsystem further comprises a noise reduction module for removing background impurity peaks in the mass spectrometry image recognition data.

[0031] Preferably, the intelligent analysis system further comprises a network visualization module for visualizing the matching results.

[0032] An intelligent analysis method for identifying dimer isomers in β-lactam antibiotics, the method comprising the following steps:

[0033] Implementation prerequisites: Actual testing and theoretical simulation of the drug to obtain mass spectrometry images and theoretical cleavage fragment sets of dimer impurities;

[0034] For each theoretical cleavage fragment, the polymerization mode is listed according to the potential polymerization site to obtain a number of dimer isomers, all potential characteristic fragments of each dimer isomer are extracted and their molecular weights are calculated to obtain corresponding potential characteristic fragment data;

[0035] Screening the potential characteristic fragment data for unique characteristic fragment data of each dimer isomer;

[0036] The mass spectrometry images obtained by the detection are screened by using the target molecular weight and the RT range in the total ion current map in which it is located, and several mass spectrometry images within the RT range are selected as target mass spectrometry images. The mass-to-charge ratio corresponding to the fragment ion peak is extracted for each target mass spectrometry image obtained by the screening to obtain the corresponding mass spectrometry image recognition data;

[0037] Each mass spectrum image recognition data is matched with each unique characteristic fragment data according to a preset fragment matching strategy to obtain a matching result.

[0038] Compared with the prior art, the present invention has the following beneficial effects:

[0039] The intelligent analysis system of the present invention constructs potential dimer isomer impurity configurations through theoretical cleavage fragments of dimer impurities, and can effectively screen out the configurations of isomer impurities by matching unique characteristic fragments of the isomer impurity configurations with mass spectrometry data, thus providing a new path for the identification of β-lactam antibiotic polymer impurities; the intelligent analysis system of the present invention can replace the manual reading and analysis of instrument data by detection personnel to perform intelligent and effective identification of dimer isomer impurities. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 Schematic diagram of the framework structure of the intelligent analysis system.

[0041] Figure 2 for Figure 1 Schematic diagram of the specific structure of the potential feature fragment data collection subsystem.

[0042] Figure 3 for Figure 1 Schematic diagram of the specific structure of the unique feature fragment data screening subsystem.

[0043] Figure 4 for Figure 1 Schematic diagram of the specific structure of the mass spectrometry image recognition subsystem.

[0044] Figure 5 for Figure 1 Schematic diagram of the specific structure of the comparison subsystem.

[0045] Figure 6 This is a diagram showing the implementation method of the network visualization module.

[0046] Figure 7 Schematic diagram of the implementation conditions and materials for the intelligent analysis method.

[0047] Figure 8 An example diagram of the relationship between potential feature fragment data and unique feature fragment data.

[0048] Fig. 9 This is an example diagram of the correspondence between the mass spectrometry image actually detected in the mass spectrometry image recognition subsystem and its mass spectrometry image recognition data.

[0049] Fig.10 This is an example diagram of the matching between the unique feature fragment data and its mass spectrometry image recognition data in the comparison subsystem.

[0050] Fig.11 It is the visualization output result of the network visualization module under different option settings. DETAILED DESCRIPTION

[0051] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. 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 belong to the protection scope of the present invention.

[0052] Reference Figure 1 As shown, an intelligent analysis system for dimer isomers in β-lactam antibiotics is composed of a potential characteristic fragment data collection subsystem, a unique characteristic fragment data screening subsystem, a mass spectrometry image recognition subsystem, a comparison subsystem, and a network visualization module.

[0053] Reference Figure 2 The potential feature fragment data collection subsystem includes a first data reading module, a feature collection module, a calculation module, and a first data output module.

[0054] The first data reading module is used to read drug information, drug sample source information, and a theoretical cleavage fragment set, where the theoretical cleavage fragment set is all possible dimer impurity cleavage fragments.

[0055] The feature acquisition module is used to arrange each theoretically understood fragment differently according to different potential polymerization sites to obtain relevant feature data of all possible dimer isomers.

[0056] The calculation module is used to develop and run an acquisition code for each dimer isomer to obtain the corresponding potential characteristic fragment data, including the molecular weight data of each potential characteristic fragment that may be generated by fragmentation in a hydrogen-containing positive ion mode.

[0057] In the feature acquisition module of the present invention, since the theoretical cleavage fragments of the dimer impurity all have multiple potential polymerization sites, the possible polymerization sites between monomers can be arranged to obtain possible dimer isomers. Taking penem drugs as an example, dimer isomers may be formed due to the carboxyl group being connected to the amino group or imino group, or the carboxyl group being connected to the hydroxyl group; taking penicillin drugs as an example, the formation of dimers may be formed by the polymerization of the carboxyl group and the amino group or imino group.

[0058] In the calculation module of the present invention, under the premise of keeping the amide or ester bond in the dimer connecting the two molecules intact, the possible dimer isomers are cut according to the cutting rules of the carbapenem compound, so that the fragments that may be generated by the breakage of the dimer isomer in the hydrogen positive ion mode can be obtained. The fragments generated by the cleavage of the same dimer isomer constitute a potential characteristic fragment set, that is, a cleavage sample. The potential characteristic fragment set under each dimer configuration will contain fragments with potential structural differences from other dimer configurations.

[0059] The first data output module is used to output potential characteristic fragment data of each dimer isomer, including fragment document code, isomers and their numbers, drug names, cleavage samples and their molecular weights (positive ions).

[0060] Reference Figure 3 As shown, the unique feature fragment data screening subsystem includes a second data reading module, a feature screening module, and a second data output module.

[0061] The second data reading module is used to read the potential characteristic fragment data of each dimer isomer, including the fragment document code, isomers and their numbers, drug names, cleavage samples and their molecular weights.

[0062] The feature screening module is used to eliminate the potential feature fragments common to other dimer isomers from the potential feature fragment data, and obtain the unique feature fragment data of each dimer isomer.

[0063] The second data output module is used to output the unique characteristic fragment data (UNIQUEprofile) of each dimer isomer, including the unique characteristic fragment document code, isomer number, drug information, cleavage sample and its molecular weight (positive ion).

[0064] Reference Figure 4 The mass spectrometry image recognition subsystem includes a third data reading module, an image recognition module, and a third data output module.

[0065] The third data reading module is used to read mass spectrometry image data, including drug name, sample conditions, mass spectrometry instrument name and model, operation mode, and actually detected LC-MS / MS mass spectrometry image.

[0066] The image data recognition and conversion module is used to recognize LC-MS / MS mass spectrometry images, and based on the target molecular weight and the RT range in the total ion current map where it is located, select several mass spectrometry images within the RT range as target mass spectrometry images, and extract the mass-to-charge ratio and mass spectrum peak intensity of any target mass spectrometry image in the positive ion mode to obtain the corresponding mass spectrometry image recognition data (MSDATA).

[0067] The total ion current diagram is a commonly used graph in mass spectrometry analysis. It shows the change in the signal intensity of all ions in the sample over time. It is an experimental data graph obtained from the actual detected LC-MS / MS mass spectrometry image.

[0068] The third data output module is used to output each mass spectrometer image recognition data, including image data document code, drug name, sample conditions, target molecular weight (positive ion), RT value and RT interval range, mass spectrometer name and model.

[0069] Reference Figure 5 The comparison subsystem includes a fourth data reading module, an accuracy control module, a comparison module, and a fourth data output module.

[0070] The fourth data reading module is used to read the unique characteristic fragment data output by the unique characteristic fragment data screening subsystem and each mass spectrum image recognition data output by the mass spectrum image recognition subsystem.

[0071] The accuracy control module is used to set a fragment matching strategy and its matching accuracy, and the fragment matching strategy includes a first matching strategy and a second matching strategy.

[0072] First matching strategy: If the absolute difference between the molecular weight (positive ion) of the unique characteristic fragment in the unique characteristic fragment data and the mass-to-charge ratio in the mass spectrum image recognition data does not exceed the preset value, the dimer isomer corresponding to the unique characteristic fragment data is matched with the mass spectrum image recognition data. The specific characterization formula of the absolute difference is:

[0073]

[0074] Wherein, D represents the difference value, m represents the charge-to-mass ratio in the mass spectrometry image recognition data under the positive ion mode, and t represents the charge-to-mass ratio of the unique characteristic fragment in the theoretical unique characteristic fragment data under the positive ion mode.

[0075] The second matching strategy: if the same unique characteristic fragment data matches no less than two mass spectrum image recognition data in the same RT interval, then it is considered that the substance in the interval corresponds to the dimer isomer of the unique characteristic fragment data.

[0076] The second matching strategy is completed on the basis of the first matching strategy. The first matching strategy is used to compare and match the unique characteristic fragment data of the dimer isomer with the mass spectrometry image recognition data of the same target molecular weight in each RT interval. After the matching is completed, the second matching strategy is executed based on the matching.

[0077] In the present invention, the preset value in the first matching strategy is the matching accuracy, which is generally between 5-20 ppm. The specific value of the matching accuracy can be set by those skilled in the art according to actual conditions.

[0078] The comparison module is used to match the theoretical characteristic fragment data of dimer isomers with the actual dimer impurity mass spectrum image recognition data based on a preset fragment matching strategy to obtain a matching result.

[0079] The fourth data output module is used to output the matching results, including the file names of the unique feature fragment data and mass spectrometry image recognition data that are successfully matched under the first matching strategy and the same numerical values ​​of the two, and / or the unique feature fragment data of the dimer isomer that are successfully matched under the second matching strategy, etc.

[0080] In the present invention, the matching accuracy is set by setting an accuracy control module, thereby improving the accuracy and specificity of the impurity risk control of β-lactam antibiotics.

[0081] Furthermore, the comparison subsystem also includes a noise reduction module for filtering out background impurity peaks with no detection significance in the mass spectrometry image recognition data.

[0082] The network visualization module is used to visualize the data based on the matching results output by the comparison subsystem, and to display the relevant data and relationship network of the unique characteristic fragments of the matched dimer isomers, such as the document name corresponding to the unique characteristic fragment, mass spectrum peak and RT value, comparison value, molecular weight in positive ion mode, cleavage sample, structural formula of the fragment, etc. According to the "Visualization Option Settings" function in the module, the display of the relationship data can be selected, such as unique data visualization, the correspondence between Excel data and unique relationship data mass spectrum data and the structural formula of the unique characteristic fragment.

[0083] The intelligent analysis system of the present invention comprises a potential characteristic fragment data collection subsystem, which performs theoretical simulation of theoretical dimer isomers and their potential characteristic fragments through code development and data intelligent processing, and a unique characteristic fragment data screening subsystem performs screening of unique characteristic fragment data of dimer isomers; and a mass spectrum image recognition subsystem extracts mass spectrum image recognition data from mass spectrum image information of impurities actually detected, and a comparison subsystem is used to achieve accurate pairing of mass spectrum image recognition data with unique characteristic fragment data, and realize multi-dimensional presentation of the relevant data relationship network.

[0084] On the basis of the above-mentioned intelligent analysis system, the present invention also provides an intelligent analysis method, which will be specifically described below taking drug a as an example.

[0085] Preconditions for implementation: Refer to Figure 7 By performing actual detection and theoretical simulation on drug a, a mass spectrometry image and all possible cleavage fragments of the dimer impurity are obtained, and all possible cleavage fragments constitute a theoretical cleavage fragment set.

[0086] Step 1, for each theoretical cleavage fragment, the polymerization mode is listed according to the potential polymerization site to obtain all possible dimer forms, and each dimer is cut according to the cutting rules of carbapenem compounds while keeping the amide or ester bond in the dimer connecting the two molecules intact to obtain the corresponding potential characteristic fragment set, and the molecular weight of each potential characteristic fragment in the positive ion mode is calculated to obtain the corresponding potential characteristic fragment data.

[0087] In step 1 of the present invention, for each theoretical cleavage fragment, different dimers can be obtained after polymerization according to its different potential polymerization sites. Different dimers of different theoretical cleavage fragments are cut respectively to obtain their corresponding potential characteristic fragment data, including the molecular weight of each potential characteristic fragment and the corresponding dimer configuration, etc. Figure 8 (a) shown.

[0088] Step 2, screening the potential characteristic fragment data of each dimer isomer to obtain the unique characteristic fragment data of each dimer isomer, such as Figure 8 (b) as shown.

[0089] In step 2 of the present invention, the potential characteristic fragments common to other dimer isomers are eliminated from the potential characteristic fragment set of each dimer isomer, and the potential characteristic fragments that are not eliminated are used as unique characteristic fragments of the dimer isomer.

[0090] Step 3, for the mass spectrum image actually detected by drug a, the target molecular weight and the RT range in the total ion current map thereof are used to screen the mass spectrum image, and at least three mass spectrum images within the RT range are selected as target mass spectrum images. Image data is extracted for each target mass spectrum image obtained by screening, such as Fig. 9 As shown in (b), the image data includes the mass-to-charge ratio and mass spectrum peak intensity in positive ion mode.

[0091] In step 3 of the present invention, the target molecular weight is the molecular weight of the dimer impurity. In the example, the molecular weight of the dimer impurity of drug a is 701. The molecular weight of the dimer impurity can be calculated based on the structure of drug a, which is common knowledge in the art. The target mass spectrum image and the mass spectrum image recognition data are in a one-to-one correspondence.

[0092] Step 4, matching the mass spectrum image recognition data of each actual dimer impurity with each theoretical unique characteristic fragment data, and matching the theoretical characteristic fragment data of the dimer isomer with the actual dimer impurity mass spectrum image recognition data according to the first matching strategy to obtain a matching result.

[0093] Fig.10 As shown, in step 4 of this embodiment, the preset value in the first matching strategy is set to 10ppm; the unique characteristic fragment data with a molecular weight of 701 is compared with each mass spectrum image recognition data, and the matching relationship between each target mass spectrum image with a mass-to-charge ratio of 701m / z in the positive ion mode and each unique characteristic fragment data is obtained, and the matching relationship is recorded in Excel, wherein column A represents the RT interval where the successfully matched target mass spectrum image is located, columns B and C are used to characterize the successfully matched target mass spectrum image and the unique characteristic fragment data, respectively, columns D and E represent the mass-to-charge ratio of the same target mass spectrum image in column B and the fragment molecular weight of the same unique characteristic fragment data in column C, respectively, and the absolute difference values ​​of the mass-to-charge ratio and the fragment molecular weight are both less than 10ppm.

[0094] Step 5: Visualize the matching results according to different options, such as Fig.11 shown.

[0095] The present invention screens out the unique characteristic fragment data obtained by theoretical calculation of each dimer isomer from the impurities actually detected and their mass spectrometry image information through code development and data intelligent processing, realizes the accurate pairing of mass spectrometry detection data with the characteristic data of potential dimers, and realizes the multi-dimensional display of the related data relationship network. The present invention provides a new path for the identification of β-lactam antibiotic polymer impurities, and also improves the accuracy and specificity of β-lactam antibiotic impurity risk control. The technical solution of the present invention is applicable to other types of drugs except β-lactam antibiotics, and can be promoted and used in the drug analysis and drug quality control industry.

Claims

1. An intelligent analysis system for identifying dimer isomers in β-lactam antibiotics, characterized in that: include: The potential characteristic fragment data collection subsystem is used to intelligently read dimer isomers and their potential characteristic fragments using theoretical cleavage fragments of dimer impurities, and obtain theoretically calculated potential characteristic fragment data of each dimer isomer, including the molecular weight of the potential characteristic fragment in positive ion mode; The unique characteristic fragment data screening subsystem is used to screen the potential characteristic fragment data of each dimer isomer to obtain the corresponding unique characteristic fragment data, including the molecular weight of the unique characteristic fragment in the positive ion mode; The mass spectrum image recognition subsystem is used to select the target mass spectrum image under the target molecular weight from the mass spectrum image obtained by the actual dimer impurity detection, and extract its mass spectrum image data, including the mass-to-charge ratio corresponding to the fragment ion peak; The comparison subsystem is used to compare the theoretically calculated unique characteristic fragment data with the actual mass spectrum image recognition data of the dimer impurity, and obtain the matching result for the molecular weight and mass-to-charge ratio according to the preset fragment matching strategy.

2. The intelligent analysis system for identifying dimer isomers in β-lactam antibiotics according to claim 1, wherein the potential characteristic fragment data collection subsystem comprises: A first data reading module is used to read theoretical cleavage fragments of dimer impurities; The feature collection module is used to list the polymerization modes based on each theoretical cleavage fragment and its different potential polymerization sites, obtain several different dimer isomers and collect relevant feature data; A calculation module, for processing relevant characteristic data for each potential characteristic fragment set of each dimer isomer, wherein the potential characteristic fragment is a fragment obtained by cutting the dimer isomer according to the cutting rule of the carbapenem compound while keeping the amide or ester bond in the dimer connecting the two molecules intact; The first data output module is used to output each dimer isomer and its potential characteristic fragment data, each potential characteristic fragment data includes the molecular weight of each potential characteristic fragment in the positive ion mode.

3. The intelligent analysis system for identifying dimer isomers in β-lactam antibiotics according to claim 1, wherein the unique characteristic fragment data screening subsystem comprises: A second data reading module is used to read potential characteristic fragment data of each dimer isomer; The feature screening module is used to eliminate the potential feature fragments common to other dimer isomers from the potential feature fragment data, so as to obtain the unique feature fragment data of each dimer isomer; The second data output module is used to output the unique characteristic fragment data of each dimer isomer, including the molecular weight of the unique characteristic fragment in the positive ion mode.

4. The intelligent analysis system for identifying dimer isomers in β-lactam antibiotics according to claim 1, characterized in that: The mass spectrum image recognition subsystem comprises: A third data reading module is used to read the mass spectrum image actually detected; The image recognition and conversion module is used to determine the RT range of the dimer impurity in the total ion current map based on the target molecular weight, obtain the mass spectrum image within the RT range as the target mass spectrum image, and extract mass spectrum image recognition data for any target mass spectrum image, including extracting the mass-to-charge ratio and mass spectrum peak intensity in the positive ion mode; The third data output module is used to output each mass spectrum image recognition data within the target RT range.

5. An intelligent analysis system for identifying dimer isomers in β-lactam antibiotics according to claim 4, characterized in that: In the image recognition module, the number of target mass spectrometry images is no less than 3.

6. An intelligent analysis system for identifying dimer isomers in β-lactam antibiotics according to claim 4, characterized in that: The comparison subsystem comprises: A fourth data reading module, used to read each unique characteristic fragment data and each mass spectrum image recognition data; The accuracy control module is used to set the fragment matching strategy and its matching accuracy, wherein the fragment matching strategy includes a first matching strategy and a second matching strategy; in the first matching strategy, when the absolute difference between the molecular weight and the mass-to-charge ratio does not exceed a preset value, the corresponding unique characteristic fragment data is determined to match the mass spectrum image recognition data; in the second matching strategy, when the same unique characteristic fragment data matches no less than two mass spectrum image recognition data in the same RT interval, the corresponding dimer isomer is determined to be consistent with the dimer impurity in the RT interval. A comparison module, for matching the molecular weight of the unique characteristic fragment data and the mass-to-charge ratio of the mass spectrum image recognition data according to a preset fragment matching strategy; The fourth output module is used to output the matching result.

7. An intelligent analysis system for identifying dimer isomers in β-lactam antibiotics according to claim 6, characterized in that: The comparison subsystem also includes a noise reduction module for removing background impurity peaks in mass spectrum image recognition data.

8. An intelligent analysis system for identifying dimer isomers in β-lactam antibiotics according to claim 1, characterized in that: The intelligent analysis system also includes a network visualization module for visualizing the matching results.

9. An intelligent analysis method for identifying dimer isomers in β-lactam antibiotics, using the intelligent analysis system according to any one of claims 1 to 8, characterized in that: The method comprises the following steps: Implementation prerequisites: Actual testing and theoretical simulation of the drug to obtain mass spectrometry images and theoretical cleavage fragment sets of dimer impurities; For each theoretical cleavage fragment, the polymerization mode is listed according to the potential polymerization site to obtain a number of dimer isomers, all potential characteristic fragments of each dimer isomer are extracted and their molecular weights are calculated to obtain corresponding potential characteristic fragment data; Screening the potential characteristic fragment data for unique characteristic fragment data of each dimer isomer; The mass spectrometry images obtained by detection are screened using the target molecular weight and the RT range in the total ion current map in which it is located, and several mass spectrometry images within the RT range are selected as target mass spectrometry images. The mass-to-charge ratio corresponding to the fragment ion peak is extracted for each target mass spectrometry image obtained by screening to obtain corresponding mass spectrometry image recognition data; each mass spectrometry image recognition data is matched with each unique characteristic fragment data according to a preset fragment matching strategy to obtain a matching result.

Images