A non-targeted screening method for veterinary drugs and their metabolites in milk
Through ultra-high performance liquid chromatography-high resolution mass spectrometry combined with mass loss filtration model, combined with food endogenous component database, the problem of difficulty in screening unknown veterinary drugs and metabolites in milk was solved, and the milk was achieved The rapid and efficient screening and identification of Chinese veterinary drugs and their metabolites has improved the ability to warn food safety risks.
Patent Information
- Application Number
- CN202110756572.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2021-07-05
- Publication Date
- 2024-11-12
- Estimated Expiration
- 2041-07-05
AI Technical Summary
Traditional targeted analysis methods are difficult to effectively screen unknown or new veterinary drugs and their metabolites in milk, resulting in a timely warning of food safety risks.
The ultra-high performance liquid chromatography-high resolution mass spectrometry combined platform collects non-targeted analysis data, combines the food endogenous component database and mass loss filtration model to remove endogenous compounds and interfering substances, and quickly screen out potential risk compounds. , and preliminary qualitative is carried out through various databases.
In the absence of negative control samples, interfering substances in milk samples are efficiently removed and risk compounds are quickly screened out, which significantly improves the ability to identify unknown veterinary drugs and their metabolites, and reduces the difficulty of identifying harmful substances.
Smart Images

Figure CN115586264B_ABST
Abstract
Description
Technical Field
[0001] The invention relates to the fields of analytical chemistry and food safety, and is a non-targeted screening method for veterinary drugs and metabolites thereof in milk. Background Art
[0002] As the quality of life improves, people pay more attention to food safety. As an important part of the diet, the nutritional value of milk can be impaired by drug residues. In particular, antimicrobial drugs are often used for postpartum treatment of dairy cows. Insufficient withdrawal period or excessive use can cause them to remain in milk, which has adverse effects on human health. Traditional targeted analysis methods focus on specific compounds, mainly targeting known veterinary drugs and common metabolites, while many veterinary drugs exist in animal-derived foods in new or unknown metabolic forms. Therefore, conducting non-targeted analysis is of great significance for screening potential unknown hazards and early warning of food safety risks.
[0003] Non-targeted analysis can obtain more compound information of the sample to be tested, but how to screen out risky compounds from complex spectral data is an urgent problem to be solved. Mass loss is the difference between the exact mass of a compound and its nominal mass, which depends on the molecular composition of the compound. It is often used to visualize related and homologous series of compounds in the environment and other fields, and is gradually being used in food safety analysis. In order to remove more interfering ions in the sample to be tested and screen out potential risk substances, a suitable mass loss filtering model can be established for non-targeted screening of veterinary drugs.
[0004] The present invention establishes a non-targeted screening method for veterinary drugs and their metabolites in milk based on food endogenous component filtration and quality loss filtration. In the absence of negative control samples, endogenous compounds and a large number of interfering substances in complex food matrices can be effectively removed, a list of potential risk compounds can be obtained, and potential risk compounds can be preliminarily characterized through multiple databases to achieve identification of unknown veterinary drugs and their metabolites. At present, there is no relevant report on the application of the method of the present invention to non-targeted screening of veterinary drugs in food. Summary of the invention
[0005] In order to realize non-targeted screening of veterinary drugs and their metabolites, the present invention collects non-targeted analysis data based on an ultra-high performance liquid chromatography-high resolution mass spectrometry platform, and simultaneously establishes a food endogenous component database and a quality defect filtering model for removing endogenous compounds and interfering substances in the sample to be tested, quickly screening out risk compounds, and finally performing preliminary qualitative analysis of risk compounds through database retrieval. The specific technical methods adopted by the present invention are as follows:
[0006] In the first step, the milk samples were pretreated using a direct organic solvent extraction method to obtain sample extraction components, and the extracts were analyzed using ultra-high performance liquid chromatography-high resolution mass spectrometry to collect non-targeted analysis data, including retention time, primary mass spectrometry, and secondary mass spectrometry information.
[0007] In the second step, the collected raw data were peak aligned and extracted to obtain the original compound list. The instrument's own software Compound Discoverer 3.1 was used to perform peak alignment and peak matching on the raw data of all samples analyzed at one time, and an Excel file of the compound list was exported, including the molecular weight, retention time, and response area in different samples. First, based on the simultaneous injection of the extractant and the resolvent with the sample, the response of the compound in the blank solvent in the exported compound list was selected, that is, the response area in at least two identical blank solvents (extractant blank or resolvent blank) was greater than 2×10 5 The remaining compounds need to be further filtered.
[0008] The third step is to remove endogenous compounds by filtering endogenous components of food. First, a database of endogenous components of food needs to be constructed. The same analytical method as the sample to be tested is used to analyze more than 80 milk samples from as many as five regions (districts at or above the county level), and non-targeted analysis data including retention time, primary mass spectrometry and secondary mass spectrometry information are collected. Then, all the data analyzed once are peak aligned and matched together, and blank interference is subtracted. According to the 80% rule, compounds detected in more than 80% of the samples are defined as endogenous components of food, so as to construct a database of endogenous components of food. In the veterinary drug screening of the sample to be tested, the molecular weight and retention time of the compound can be compared with the compounds in the database of endogenous components of food. Compounds with a molecular weight deviation within 5ppm and a retention time deviation within 0.2min are considered to be the same compounds, so as to remove the same compounds as those in the database of endogenous components of food.
[0009] In the fourth step, the mass loss filtering method is used to further remove interfering substances and screen out potential risk compounds. First, more than 80 (preferably more than 100) veterinary drugs in five or more categories (preferably more than seven categories) are selected as screening drugs, and new veterinary drugs and veterinary drug metabolites belonging to this category are screened non-targeted, and used to construct a mass loss filtering model. Secondly, the constructed mass loss filtering model is used to remove interfering substances in experimental samples.
[0010] (1) Using the drawing software OriginPro 2018, the veterinary drugs were displayed in a rectangular coordinate system, with the nominal mass of the veterinary drug (the nominal mass refers to the integer value of the sum of the masses of various atoms constituting a certain ion or particle, which is expressed as the atomic weight of 12C being 12, and this unit is used to represent the integer value, that is, the integer part of the accurate molecular weight) and the mass loss of the veterinary drug (the mass loss refers to the difference between the accurate mass of an element or compound and its nearest integer value, that is, the decimal part of the accurate molecular weight, where the accurate molecular weight is accurate to 0.001Da) as the x and y axes, and then the corresponding coordinate points of the veterinary drug were surrounded by rectangles, parallelograms and polygons respectively. Specifically, according to the needs of non-targeted screening, rectangular and parallelogram mass loss filtering models were established by moving 50–100 mDa, or 30 mDa, in the positive and negative directions of the y-axis in the mass loss distribution area of veterinary drugs. At the same time, a polygonal mass loss filtering model based on the distribution area of veterinary drugs and their metabolites was proposed. The veterinary drug metabolites were predicted using Meteor Nexus 3.1 software to provide rich metabolite information, and the coordinate points corresponding to the veterinary drugs and the predicted veterinary drug metabolites were surrounded by fine polygons to construct a polygonal mass loss filtering model based on the distribution area of veterinary drugs and their metabolites.
[0011] (2) In the veterinary drug screening of the samples to be tested, the coordinate points corresponding to the compounds in the samples are also displayed in the mass loss diagram. The compounds falling within the boundary of the mass loss filtering model and the boundary are risk compounds, and the compounds falling outside the model are removed interfering substances. The interfering substances are further removed through the mass loss filtering model.
[0012] Step 5: Database search for preliminary characterization of risk compounds. Database search, including DrugBank, MassBank, and Metlin databases, was performed to identify veterinary drugs and their metabolites in risk compounds.
[0013] The present invention uses an ultra-high performance liquid chromatography-high resolution mass spectrometry platform to obtain non-targeted analysis data for veterinary drugs and their metabolites in milk, removes a large number of endogenous compounds and interfering substances through filtering of endogenous components of food and filtering of multiple mass loss models, and realizes rapid screening of veterinary drugs and their metabolites in milk samples without negative control samples. The invention has significant advantages in non-targeted screening of veterinary drugs and their metabolites in milk samples without prior information, and has the advantages of efficiently removing interfering substances and rapidly screening risk compounds. BRIEF DESCRIPTION OF THE DRAWINGS
[0014] Figure 1Three mass loss filtering models for filtering interfering substances. (A) Rectangular mass loss filtering model, (B) Parallelogram mass loss filtering model, (C) Polygonal mass loss filtering model based on the distribution area of veterinary drugs and their metabolites, (D)(E)(F) are three mass loss filtering models used for filtering interfering substances in the milk samples to be tested. The black data points are the removed interfering substances, and the red data points are the risk compounds.
[0015] Figure 2 Examples of non-targeted screening compounds in samples. (A) N-desmethyl ofloxacin in sample, (B) N-desmethyl ofloxacin standard, (C) Ciprofloxacin-M3 in sample, (D) Ciprofloxacin-M3 standard. DETAILED DESCRIPTION
[0016] The following is a detailed description of the implementation of the present invention in conjunction with the attached tables and drawings: This embodiment is implemented on the premise of the technical solution of the present invention, and a detailed implementation method and specific operation process are given, but the protection scope of the present invention is not limited to the following embodiments.
[0017] Example 1
[0018] Sample information:
[0019] 134 milk samples (samples for establishing the food endogenous component database) were collected from individual dairy farmers in five different cities, namely Harbin, Qingdao, Hulunbuir, Hohhot and Dalian, for the establishment of the food endogenous component database, including 25 samples from Harbin, 37 samples from Qingdao, 33 samples from Hulunbuir, 24 samples from Hohhot and 15 samples from Dalian.
[0020] Afterwards, another 36 milk samples (test samples) were randomly collected from the local market in Hulunbuir for veterinary drug screening experiments. All samples were freeze-dried and stored in a refrigerator at -40°C for the next step of analysis.
[0021] 1. Non-targeted data acquisition based on ultra-performance liquid chromatography-high-resolution mass spectrometry.
[0022] Sample analysis: Sample pretreatment was performed by direct solvent extraction. Specifically, 100 mg (±1 mg) of lyophilized milk sample was weighed into a 2 mL centrifuge tube, 300 μL of water was added, and vortexed for 2 minutes to completely dissolve it. Then, 1 mL of acetonitrile (1 / 99, v / v) containing 1% formic acid was added and vortexed for 2 minutes. Then, the sample was centrifuged at 21500 g and 4 °C for 10 minutes in a high-speed centrifuge (Hitachi, Ibaraki, Japan), 800 μL of supernatant was taken into another 2 mL centrifuge tube, and freeze-dried in a centrifugal vacuum freeze dryer (Labconco, Kansas, USA). The lyophilized sample was reconstituted with 200 μL of 20% acetonitrile (v / v) resolvent, vortexed for 1 minute, and then centrifuged at 21500 g and 4 °C for 10 minutes in a high-speed centrifuge. The supernatant was taken into another 2 mL centrifuge tube and centrifuged again. Finally, the supernatant was transferred to an injection vial, and the samples were analyzed by simultaneous acquisition of primary and secondary mass spectrometric data using ultra-high performance liquid chromatography (UHPLC, Waters, Milford, MA, USA)-high-resolution mass spectrometry (Q Exactive HF, Thermo Fisher Scientific, Rockford, IL, USA) technology.
[0023] 2. Screening of risk compounds in samples.
[0024] In this example, the method in step (1) above was used to analyze 36 milk samples (samples to be tested). After sample pretreatment and ultra-high performance liquid chromatography-high resolution mass spectrometry analysis, non-targeted data including primary and secondary mass spectra were obtained. At the same time, the extractant blank (6 samples, 180 μL each) and the resolvent blank (6 samples, 180 μL each) were injected and analyzed separately. The original data of all samples analyzed at one time were peak matched together using the instrument's own software Compound Discoverer 3.1 to obtain a compound list including the compound molecular weight, retention time, and response area in different samples. First, based on the response of the compound in the blank solvent in the derived compound list, that is, the response area in at least two identical blank solvents (extractant blank or resolvent blank) was greater than 2×10 5 The substances were deleted, and 3122 compounds remained in the compound list, which were then filtered by endogenous components of food and quality loss model to remove endogenous compounds and interfering substances and screen out risky compounds.
[0025] The first step is to filter the endogenous components of food. First, a database of endogenous components of food is constructed. 134 milk samples from different sources (samples for establishing the database of endogenous components of food) are analyzed using the same analytical method as the samples to be tested to obtain non-targeted data including primary and secondary mass spectra. The original data is peak extracted and blank interference is subtracted. According to the 80% rule, that is, the compounds detected in more than 80% of the samples are the required compounds, the endogenous small molecule composition of food is obtained, and a database of endogenous components of food containing 2237 compounds is constructed, including the molecular weight and retention time of the compounds.
[0026] In the veterinary drug screening of 36 samples to be tested, the remaining 3122 compounds were compared with the molecular weight and retention time of the compounds in the food endogenous component database. The compounds with molecular weight deviation within 5 ppm and retention time deviation within 0.2 min were considered to be the same compounds, so as to remove the compounds identical to those in the food endogenous component database. Among them, 1738 endogenous compounds were removed, and 1384 compounds remained.
[0027] The second step is mass loss model filtering. Antimicrobial drugs are often used for treatment and prevention during dairy cow production, so 112 commercially available antimicrobial drugs belonging to seven categories of antimicrobial drugs were selected, including sulfonamides, quinolones, penicillins, tetracyclines, benzimidazoles, nitrofurans and nitroimidazoles, and a mass loss filtering model was established to remove interfering substances from the remaining 1384 compounds. Detailed information on the 112 veterinary drugs is shown in Table 1. Three mass loss filtering models and their effects on interfering substance filtering are shown in Figure 1 .
[0028] (1) Construct three quality loss filtering models. Using the drawing software OriginPro 2018, 112 veterinary drugs were displayed in a plane rectangular coordinate system, with the nominal mass of the veterinary drugs (nominal mass refers to the set 12 The atomic weight of C is 12, and this unit is used to represent the integer value of the sum of the masses of various atoms that make up a certain ion or particle, that is, the integer part of the accurate molecular weight) as the x-axis, the mass loss of veterinary drugs (mass loss refers to the difference between the accurate mass of an element or compound and its nearest integer value, that is, the decimal part of the accurate molecular weight, where the accurate molecular weight is accurate to 0.001Da) as the y-axis, and then rectangles, parallelograms and polygons are used to surround the corresponding coordinate points of the veterinary drugs to establish a mass loss filtering model.
[0029] Specifically, 112 veterinary drugs are used to establish a rectangular mass loss filtering model and a parallelogram mass loss filtering model. The boundaries of the rectangular mass loss filtering model are the maximum nominal mass (neobiocin: 613Da) and the minimum nominal mass (metronidazole: 128Da) of the 112 veterinary drugs, which are moved 50Da in the positive and negative directions of the x-axis, that is, the maximum nominal mass +50Da, the minimum nominal mass -50Da, which are the x-axis coordinates of the four vertices of the rectangle respectively; the maximum mass loss (neobiocin: 239.2mDa) and the minimum mass loss (triclabendazole: -42.6mDa) are moved 50mD in the positive and negative directions of the y-axis, that is, the maximum mass loss +50mDa, the minimum mass loss -50mDa, which are the y-axis coordinates of the four vertices of the rectangle respectively. Therefore, the four vertices of the rectangle are (78Da, -92.6mDa), (78Da, 289.2mDa), (663Da, -92.6mDa), (663Da, 289.2mDa). The parallelogram mass loss filtering model takes into account that the mass loss of a compound is determined by the number of hydrogen atoms, and the mass loss increases with the molecular weight of the compound. A parallelogram with the smallest area is drawn and can surround all the coordinate points corresponding to 112 veterinary drugs. Then the vertices of the parallelogram move 50Da in the positive and negative directions of the x-axis, that is, the maximum nominal mass +50Da, the minimum nominal mass -50Da, which are the x-axis coordinates of the four vertices of the parallelogram; and move 30mDa in the positive and negative directions of the y-axis, that is, the maximum mass loss +30mDa, the minimum mass loss -30mDa, which are the y-axis coordinates of the four vertices of the parallelogram. Therefore, the four vertices of the parallelogram are (78Da, -234.7mDa), (78Da, 89.8mDa), (663Da, 81.4mDa), (663Da, 405.9mDa), and a parallelogram mass loss filtering model with a smaller interval than the rectangular mass loss filtering model is constructed. At the same time, a polygonal mass loss filtering model based on the distribution area of veterinary drugs and their metabolites is proposed. First, the veterinary drug metabolites were predicted using Meteor Nexus 3.1 software. The veterinary drug structure was imported into the software, and the software automatically predicted the veterinary drug metabolites. The excel file was exported, including the molecular weight, structure and other information of the veterinary drug metabolites. The metabolites of 112 veterinary drugs were predicted, totaling 900 metabolites. The coordinate points corresponding to the 1012 veterinary drugs and the predicted metabolites were displayed in the mass loss diagram, which was distributed in an irregular shape. All the coordinate points located at the boundary of the figure (the number of coordinate points was 28) were connected end to end with line segments, and the closed polyline formed surrounded all the coordinate points. The area surrounded by the closed polyline represents the distribution area of veterinary drugs and veterinary drug metabolites. Since the measurement error of the instrument is inevitable, the closed polyline was moved 3mDa outward along the normal direction of the straight line to form a larger range that can accommodate errors, so as to construct a polygonal mass loss filtering model with a fine contour.The veterinary drug classes and their molecular weight and mass loss distribution intervals, as well as the number of veterinary drugs in each class and the predicted metabolites are shown in Table 2.
[0030] (2) Three mass defect filtering models were used for interfering substance filtering. The coordinate points corresponding to 1384 compounds were displayed simultaneously in the same plane rectangular coordinate system with the three mass defect filtering models established above using the drawing software OriginPro 2018. Compounds falling within the boundary of the mass defect filtering model and the boundary were risk compounds, while compounds falling outside the mass defect filtering model were removed interfering substances. The rectangular mass defect filtering model removed 441 (32%) compounds and screened out 943 risk compounds. The parallelogram mass defect filtering model removed 594 (43%) compounds and screened out 790 risk compounds. The polygonal mass defect filtering model removed 1058 (76%) compounds and screened out 326 risk compounds. Compared with the rectangular and parallelogram mass defect filtering models established directly using veterinary drugs, the polygonal mass defect filtering model based on the distribution area of veterinary drugs and predicted metabolites has a more significant filtering effect. Finally, 326 risk compounds were screened out from 3122 compounds in the original compound list through filtering of endogenous components in food and polygon mass loss.
[0031] 3. Database retrieval is used for preliminary identification of risk compounds.
[0032] The databases include DrugBank database, MassBank database, and Metlin database, which are used to preliminarily identify the 326 risk compounds screened out after the above two steps of filtering. Ten veterinary drugs and veterinary drug metabolites were preliminarily identified, including 4 veterinary drugs and 6 veterinary drug metabolites. The veterinary drugs are enrofloxacin (m / z 360.1718), ofloxacin (m / z 362.1511), oxytetracycline (426.1181) and levamisole (m / z 205.0797), and the veterinary drug metabolites are ciprofloxacin (m / z 332.1404), N-desmethylofloxacin (m / z 348.1353), ciprofloxacin-M3 (m / z 346.1198), as well as desethylene ciprofloxacin (m / z 306.1247), formyldesethylene ciprofloxacin (m / z 334.1193) and desethylene enrofloxacin (m / z 334.1556). The structures of 7 compounds have been confirmed by standard samples. The structures of compounds m / z 306.1247, m / z 334.1193 and m / z 334.1556 were inferred by mass spectrometry because there were no commercial standards. The veterinary drug metabolites N-desmethylofloxacin and ciprofloxacin-M3 were verified by purchasing standard samples. The qualitative results are shown in Figure 2 .
[0033] The screened veterinary drugs and their metabolites were quantitatively analyzed. The screened veterinary drugs and their metabolite standards were added to blank milk samples in different amounts to form a series of concentration gradients (0.1, 0.2, 0.4, 1, 2, 4, 10, 20, 40, 100, 200, 400, 1000, 2000 μg / kg) and analyzed under the same operating conditions as the actual samples containing veterinary drugs. A standard curve was drawn based on the concentration and response of the standard in the blank milk matrix to achieve the quantification of veterinary drugs and their metabolites in the actual samples. The quantitative results are shown in Table 3.
[0034] The present invention targets milk samples, utilizes an ultra-high performance liquid chromatography-high resolution mass spectrometry platform to obtain non-targeted analysis data, and adopts food endogenous component filtering and mass loss model filtering methods to remove a large number of endogenous compounds and interfering substances, discover potential risk compounds, and greatly reduce the difficulty of identifying harmful substances. It has the advantages of efficiently removing interfering substances and quickly screening risk compounds.
[0035] Table 1. Detailed information of 112 veterinary drugs in seven categories used to establish the quality defect filtering model
[0036]
[0037]
[0038]
[0039]
[0040]
[0041]
[0042] Table 2. Veterinary drug classes, molecular weights, and mass loss distribution intervals, as well as the number of veterinary drugs and predicted metabolites for each class
[0043]
[0044]
[0045] Table 3. Quantitative results of veterinary drugs and their metabolites screened in milk samples (unit: μg / kg)
[0046]
Claims
1. A non-targeted screening method for veterinary drugs and their metabolites in milk, characterized by: The method comprises the following steps: (1) The milk sample to be tested is pretreated to obtain sample extraction components, and the extract is analyzed by ultra-high performance liquid chromatography-high resolution mass spectrometry to obtain non-targeted analysis data; (2) Screening of risky compounds: Peak alignment and peak extraction are performed on non-targeted analysis data to obtain the original compound list. After deducting blank interference, risky compounds are screened out through food endogenous component filtering and mass loss model filtering; (3) Database retrieval is used to characterize risk compounds and preliminarily identify veterinary drugs and their metabolites; In step (1), a direct organic solvent extraction pretreatment method is required to extract the veterinary drug small molecule components from the milk sample into the extractant. In step (1), the sample is analyzed using a method of simultaneously collecting primary and secondary mass spectrometric data using ultra-high performance liquid chromatography-high resolution mass spectrometry technology to obtain original spectroscopic data, including retention time, primary mass spectrum and secondary mass spectrum information. To obtain the compound list in step (2), the software provided by the ultra-high performance liquid chromatography-high resolution mass spectrometry instrument is used to simultaneously perform peak alignment and peak matching on the raw data of all samples analyzed at one time. The exported compound list file contains the compound molecular weight, retention time, and response area in different samples. In step (2), the method of filtering endogenous components of food is used to screen risk compounds. The specific process is as follows: more than 80 milk samples from 5 county-level and above regions are analyzed in the same way as in step (1). After peak matching and blank subtraction, the compounds detected in more than 80% of the samples are defined as endogenous components of food according to the 80% rule. In this way, a database of endogenous components of food is established, which includes the molecular weight and retention time of the compounds. Filtering of endogenous components in food: In the veterinary drug screening of milk samples to be tested, the molecular weight and retention time of the compounds in the samples are compared with the compounds in the endogenous component database of food. The compounds with a molecular weight deviation within 5 ppm and a retention time deviation within 0.2 min are considered to be the same compounds, so as to remove the same compounds as those in the endogenous component database of food; The method of using the mass defect filtering model to screen risky compounds in step (2) is as follows: (A) more than 80 veterinary drugs in five categories or more are selected as screening drugs, and new veterinary drugs and veterinary drug metabolites belonging to these categories are screened by non-targeted screening, and used to construct a mass loss filtering model; the veterinary drugs are displayed in a plane rectangular coordinate system using drawing software, with the nominal mass of the veterinary drugs as the x-axis and the mass loss of the veterinary drugs as the y-axis, and then the coordinate points corresponding to the veterinary drugs are surrounded by one, two or three of a rectangle, a parallelogram and a polygon, respectively, to construct a mass loss filtering model; the nominal mass refers to the set 12 The atomic weight of C is 12, and this unit is used to represent the integer value of the sum of the masses of various atoms constituting a certain ion or particle, which is the integer part of the accurate molecular weight. The mass defect refers to the difference between the accurate mass of an element or compound and its nearest integer value, which is the decimal part of the accurate molecular weight, where the accurate molecular weight is accurate to 0.001 Da; Specifically, the boundaries of the rectangular mass loss filtering model are the maximum nominal mass and the minimum nominal mass in the veterinary drug shifted by 50–100 Da in the positive and negative directions of the x-axis, that is, the maximum nominal mass +(50–100) Da, the minimum nominal mass -(50–100) Da, which are the x-axis coordinates of the four vertices of the rectangle respectively; the maximum mass loss and the minimum mass loss shifted by 50–100 mDa in the positive and negative directions of the y-axis, that is, the maximum mass loss +(50–100) mDa, the minimum mass loss -(50–100) mDa, which are the y-axis coordinates of the four vertices of the rectangle respectively; The parallelogram mass loss filtering model takes into account that the mass loss of a compound is determined by the number of hydrogen atoms, and the mass loss increases with the increase of the molecular weight of the compound. A parallelogram with the smallest area is drawn and can surround all the coordinate points corresponding to the veterinary drug. Then, the vertices of the parallelogram are moved by 50–100 Da in the positive and negative directions of the x-axis, that is, the maximum nominal mass +(50–100) Da, the minimum nominal mass -(50–100) Da, which are the x-axis coordinates of the four vertices of the parallelogram respectively; and moved by 30–50 mDa in the positive and negative directions of the y-axis, that is, the maximum mass loss +(30–50) mDa, the minimum mass loss -(30–50) mDa, which are the y-axis coordinates of the four vertices of the parallelogram respectively, thereby constructing a parallelogram mass loss filtering model with a smaller range than the rectangular mass loss filtering model. Based on the polygonal mass loss filtering model of the distribution area of veterinary drugs and their metabolites, the veterinary drug metabolites were first predicted using Meteor Nexus3.1 software, the structure of the veterinary drug was imported into the software, the software automatically predicted the veterinary drug metabolites, and the molecular weight and structural information of the veterinary drug metabolites were exported in an Excel file. Then, the coordinate points corresponding to the veterinary drug and the predicted veterinary drug metabolites were displayed in the mass loss diagram, which was distributed in an irregular shape. All the coordinate points located at the boundary of the figure were connected end to end with line segments, and the closed polyline formed surrounded all the coordinate points; the area surrounded by the closed polyline represented the distribution area of veterinary drugs and veterinary drug metabolites; because the measurement error of the instrument is inevitable, the closed polyline was moved outward along the normal direction of the straight line by 2-5 mDa to form a larger range of tolerance errors, so as to construct a polygonal mass loss filtering model with a fine contour; (B) In the veterinary drug screening of the samples to be tested, the coordinate points corresponding to the compounds in the samples are also displayed in the mass loss diagram. The compounds falling within the boundary of the mass loss filtering model and the boundary are risk compounds, and the compounds falling outside the mass loss filtering model are removed interfering substances, and the interfering substances are further removed through the mass loss filtering model.
2. The method according to claim 1, characterized in that: In step (1), the extractant is acetonitrile containing 1-2% by volume formic acid, and the resolvent is acetonitrile having a volume concentration of 20-50%.
3. The method according to claim 1, characterized in that: The blank interference is deducted in step (2) based on the same operating conditions for the sample and more than two extractants and more than two complex solvents, and the peak matching of the sample and the blank solvent is performed together. According to the response of the compound in the blank solvent in the derived compound list, the response area of the compound in at least two identical blank solvents is greater than 2×10 5 The substance is deleted.
4. The method according to claim 1, characterized in that: The database in step (3) is one or more of the network databases, including one or more of the DrugBank database, the MassBank database, and the Metlin database.
