Chemical indicator fingerprint spectrum for identifying effluent pollution of sewage treatment plant and textile printing and dyeing mill and application of chemical indicator fingerprint spectrum

Abstract

The invention relates to a chemical indicator fingerprint spectrum for identifying effluent pollution of a sewage treatment plant and a textile printing and dyeing mill and application of the chemical indicator fingerprint spectrum, and belongs to the technical field of water environment industrial point source pollution source analysis. The chemical indicator fingerprint spectrum is composed of 28 compounds including benzene rings, heterocyclic rings, organic acids, derivatives thereof and the like. The chemical indicator fingerprint spectrum is only used for identifying pollution of a sewage treatment plant and pollution of a textile printing and dyeing mill. The 28 characteristic chemical indicators screened by the method can quickly and accurately distinguish and identify the pollution of the sewage treatment plant and the pollution of the textile printing and dyeing mill, the evaluation index accuracy is 0.9333, the AUC is 0.9500, the recall rate is 0.9333, the precision is 0.9667, and the F1 is 0.9333. The method has the advantage that the accuracy and reliability of identifying the pollution of the sewage treatment plant and the pollution of the textile printing and dyeing mill are improved.

Description

Technical Field

[0001] The invention belongs to the technical field of water environment industrial point source pollution source analysis, and particularly relates to a chemical indicator fingerprint spectrum and application thereof for identifying effluent pollution from sewage treatment plants and textile printing and dyeing plants. Background Art

[0002] With the continuous acceleration of industrialization, the discharge of industrial wastewater is increasing day by day, and the complex pollutants contained in it pose a serious threat to the aquatic environment ecosystem. Traditional environmental monitoring methods usually rely on the pre-identification of target pollutants, which is difficult to fully reflect the true situation of pollutants in industrial wastewater, and limits the ability to identify new pollutants and unknown risks.

[0003] In recent years, non-targeted screening technology has been increasingly widely used in the field of environmental analysis due to its advantages of high throughput, high sensitivity, and no need to predetermine target pollutants. This technology, combined with analytical methods such as high-resolution mass spectrometry and liquid chromatography, can simultaneously detect and identify a large number of compounds in environmental samples, and then identify potential characteristic pollutants. Compared with traditional targeted analysis methods, non-targeted screening technology can fully understand the composition of pollutants in industrial wastewater, help discover new environmental risk substances, and provide a more scientific basis for the prevention and control of water environmental pollution. At present, non-targeted screening technology has been used to identify characteristic pollutants in industrial wastewater in the pharmaceutical, printing and dyeing, chemical and other industries, and combined with chemometric methods such as principal component analysis and cluster analysis, the source of water environmental pollution has been preliminarily analyzed, and the traceability and identification of industrial point source pollution has been preliminarily realized. However, in the face of challenges such as the complexity of massive data processing and analysis, the intervention of machine learning can help solve this problem. Based on a large number of samples from a certain industry, a pollutant fingerprint spectrum of a single industry is constructed to help trace the source. Summary of the invention

[0004] To this end, the technical problem solved by the present invention is: to provide a chemical indicator fingerprint spectrum and its application for identifying effluent pollution from sewage treatment plants and textile printing and dyeing factories; by detecting the species composition of compounds in water bodies and combining with a secondary discriminant analysis model, it is possible to quickly and accurately distinguish whether the pollution comes from effluent from sewage treatment plants or textile printing and dyeing factories.

[0005] To this end, the first technical solution of the present application discloses a chemical indicator fingerprint spectrum for identifying effluent pollution from sewage treatment plants and textile printing and dyeing plants, wherein the fingerprint spectrum is obtained by screening characteristic components of the target water body; the screening includes characteristic component extraction and characteristic component selection; wherein the characteristic component extraction includes: Pre-treat the target water sample; The compounds in the pretreated target water samples were extracted by connecting Oasis HLB, ISOLUTE ENV+, Strata XCW, and Strata XAW solid phase extraction cartridges in series; The extracted characteristic components are analyzed by mass spectrometry and matched with the database to obtain a characteristic component data set.

[0006] Furthermore, the chromatographic column used in the mass spectrometry analysis is ACQUITY UPLC BEH C18, and the pre-protection column is: ACQUITY UPLC BEH C18.

[0007] Furthermore, the characteristic components are selected by machine learning, specifically: After the characteristic component data set is classified, the importance of each feature is determined by combining the Boruta-SHAP algorithm with the SHAP value to obtain characteristic pollutants; Construct classification prediction models and evaluate models for characteristic pollutants.

[0008] Furthermore, a classification prediction model is constructed based on the quadratic discriminant analysis algorithm.

[0009] Furthermore, the model evaluation indicators include classification accuracy, area under the curve, recall rate, and accuracy.

[0010] Furthermore, the chemical indicator fingerprint spectrum includes 28 characteristic components, among which the components used to identify sewage treatment plants include pazufloxacin mesylate, octyl methoxycinnamate, oleic acid amide, demethyl-formamido-pirimicarb, octocrylene, (4-nonylphenoxy)-acetic acid, propoxur, and benzylsulfuron-methyl; the components used to identify textile printing and dyeing factories include bensulfuron, pirimicarb, propiconazole, ritalin, olmesartan, butazone, dimethoate, ethoxyquin, TCPP, triadimenol, 4-formamidoantipyrine, sucralose, estriol, neostigmine, climbazole, isopropylamine, N-butylbenzenesulfonamide, sulfadiazine, and tapentadol.

[0011] The second technical solution of the present application discloses the application of the above-mentioned chemical indicator fingerprint spectrum in identifying that the source of water pollution belongs to sewage treatment plants and / or textile printing and dyeing factories.

[0012] Furthermore, the application comprises the following steps: (1) Extract information about compounds contained in target water samples; (2) Identify the compound information based on the quadratic discriminant analysis model to determine the source of the target water sample.

[0013] Preferably, the step (1) is solid phase extraction, specifically extraction by connecting Oasis HLB, ISOLUTE ENV+, Strata XCW, and Strata XAW solid phase extraction columns in series.

[0014] Compared with the existing traceability technology, the present invention has the following beneficial effects: 1. High-throughput non-targeted screening technology: Use four types of complementary SPE columns and LC / Q-TOF in series for non-targeted screening, retaining as many molecules as possible to obtain richer molecular information.

[0015] 2. The Boruta-SHAP feature selection method combines the tree-based feature importance with SHAP to rank all features and use statistical tests to determine the importance of each feature, thereby screening out the compounds that most affect the classification results.

[0016] 3. The selection of feature selectors is based on machine learning of a large number of samples, taking into account the different expressions of compounds in the same type of samples and the high and low expressions of samples in different categories. The machine learning model established by this method has high generalization ability and stability. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 The features are arranged in descending order of importance after the Boruta-SHAP feature selection algorithm; Figure 2 The relative abundance (mass spectral intensity) of each compound in all samples of the two categories before feature selection; Figure 3 is the relative abundance (mass spectrum intensity) of each compound in all samples of the two categories after feature selection; Figure 4 Evaluation of various parameters of the quadratic discriminant analysis model in classification prediction before and after feature selection. DETAILED DESCRIPTION

[0018] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions 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 are within the scope of protection of the present invention.

[0019] Unless otherwise specifically stated, the relative arrangement, numerical expressions and numerical values ​​of the parts and steps set forth in these embodiments do not limit the scope of the present application. At the same time, it should be understood that, for ease of description, the sizes of the various parts shown in the accompanying drawings are not drawn according to the actual proportional relationship. The technology, method and equipment known to those of ordinary skill in the relevant field may not be discussed in detail, but in appropriate cases, the technology, method and equipment should be considered as a part of the authorization specification. In all examples shown and discussed here, any specific value should be interpreted as being merely exemplary, rather than as a limitation. Therefore, other examples of exemplary embodiments may have different values. It should be noted that similar reference numerals and letters represent similar items in the following drawings, so that once a certain item is defined in an accompanying drawing, it does not need to be further discussed in subsequent drawings.

[0020] The experimental methods in the following examples are all conventional methods unless otherwise specified. The experimental materials used in the following examples were purchased from commercial channels, and the purity of the reagents was basically LC-MS grade or LC grade. At the same time, quality control measures were introduced into the experimental process to ensure the quality of the data results.

[0021] Example 1 Construction of a screening-level classification prediction model for identifying chemical indicator fingerprints of effluent pollution from sewage treatment plants and textile printing and dyeing plants S1. Sample collection: 500 ml of effluent from various factories was collected for each independent sample. 20 samples from sewage treatment plants and 22 samples from textile printing and dyeing factories were stored at 4°C and transported back to the laboratory, and the samples were pretreated within 24 hours.

[0022] S2. Sample pretreatment: The water sample was filtered using a 0.45 μm glass fiber filter.

[0023] S3. Solid phase extraction treatment: adjust the pH of the water sample to about 6.5, select Oasis HLB (200 mg), ISOLUTE ENV+ (150 mg), Strata XCW (100 mg), Strata XAW (100 mg) solid phase extraction columns in series for compound extraction, load the sample at a rate of 10 ml / min, and elute step by step with 6 ml of ethyl acetate:methanol (1:1) solution containing 0.5% ammonia and 3 ml of ethyl acetate:3 ml methanol (1:1) solution containing 1.7% formic acid; concentrate to 100 μL under nitrogen conditions and re-adjust to 1 mL, and store in a brown ampoule for instrument analysis.

[0024] S4. Instrumental analysis: The positive ion mode of Agilent 6546 LC / Q-TOF high-resolution mass spectrometer was selected for non-targeted analysis. The chromatographic column used was ACQUITY UPLC BEH C18 (1.7 μm, 2.1×100 mm), and the pre-guard column was ACQUITY UPLC BEH C18 (1.7 μm VANGUARD). Other instrument settings are shown in Tables 1 and 2.

[0025] Table 1 Agilent 6546 LC / Q-TOF high-resolution mass spectrometer non-targeted analysis instrument settings

[0026] Table 2 Mobile phase gradient settings

[0027] S5. Mass spectrometry batch peak extraction and compound identification: The Mass Profinder software of the Agilent platform was used for batch extraction of molecular features. The Mass Profiler Professional software was used for data set preprocessing and compound library identification. The databases used were Agilent MassHunter Water Screening PCDL and AgilentMassHunter EPA PCDL for matching. Compounds with matching scores greater than 90 were retained to obtain the data set.

[0028] S6. Compound batch structural classification and annotation: Obtain the compound list and its InChIKey identifiers, select ClassyFire to automatically annotate and structurally classify the InChIKey identifiers of the compounds, and use the compound's Superclass to classify the compounds.

[0029] S7. Machine learning feature selection: The feature selection algorithm selects the Boruta-SHAP algorithm, which combines the classic Boruta algorithm with the SHAP (SHapley Additive exPlanations) value based on the random forest model, sorts all features and uses statistical tests to determine the importance of each feature (such as Figure 1 ), Figure 1 Accepted means that the Boruta-SHAP algorithm is used to calculate the features of each variable that contribute most to the prediction results of the classification model, thereby screening out the compounds that best reflect the classification results, namely, characteristic pollutants (as shown in Table 3).

[0030] Table 3 Names of 28 important features after feature selection, corresponding InChIKey and their categories

[0031]

[0032] Thus, 28 important feature names were screened out as chemical indicator fingerprints for identifying effluent pollution from sewage treatment plants and / or textile printing and dyeing factories.

[0033] S8. Classification prediction model construction: Based on the quadratic discriminant analysis algorithm (Quadratic Discriminant Analysis Algorithm) classification model (model parameter settings: 10-fold cross validation, 7:3 division of training set and validation set, random number 42), the data set obtained in step S5 is used to repeat the characteristic pollutants obtained in step S7 above to build the model. It should be noted that this model is only suitable for distinguishing whether the sample category is a sewage treatment plant or a textile printing and dyeing factory. The data distribution before and after feature selection is as follows Figure 2 , 3 .

[0034] S9. Model evaluation: The following four indicators including ROC classification accuracy (Accuracy), area under the curve (AUC), classification accuracy (Accuracy), recall (Recall), accuracy (Precision), F1, Kappa, and MCC were used to evaluate the classification model. The results are as follows Figure 4 As shown in the figure, the results show that after using 28 source characteristic chemical indicators for QDA model training, 20 sewage treatment plant samples and 22 textile printing and dyeing plant samples can be significantly distinguished from the test set of 42 pollution source samples, with a classification accuracy of 0.9333, an AUC of 0.9500, a recall rate of 0.9333, a precision of 0.9667, and an F1 of 0.9333. All the results show that the characteristic pollutants screened by the features can be used to construct a chemical indicator fingerprint library that can significantly distinguish sewage treatment plants and textile printing and dyeing plants.

[0035] The above embodiments are only used to illustrate the technical solutions of the present invention, but the protection scope of the present invention is not limited thereto. The embodiments of the present invention and the features in the embodiments can be arbitrarily combined with each other without conflict. Any technician familiar with the technical field can easily think of changes or replacements within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A chemical indicator fingerprint for identifying effluent pollution from sewage treatment plants and textile printing and dyeing plants, characterized in that: The fingerprint spectrum is obtained by screening the characteristic components of the target water body; The screening includes characteristic component extraction and characteristic component selection; wherein the characteristic component extraction includes: Pre-treat the target water sample; The compounds in the pretreated target water samples were extracted by connecting Oasis HLB, ISOLUTE ENV+, Strata XCW, and Strata XAW solid phase extraction cartridges in series; The extracted characteristic components are analyzed by mass spectrometry and matched with the database to obtain a characteristic component data set.

2. The chemical indicator fingerprint spectrum according to claim 1, characterized in that: The chromatographic column used for the mass spectrometry analysis is ACQUITY UPLC BEH C18, and the pre-protection column is: ACQUITY UPLC BEH C18.

3. The chemical indicator fingerprint spectrum according to claim 1, characterized in that: The feature component selection is performed through machine learning, specifically: After the characteristic component data set is classified, the importance of each feature is determined by combining the Boruta-SHAP algorithm with the SHAP value to obtain characteristic pollutants; Construct classification prediction models and evaluate models for characteristic pollutants.

4. The chemical indicator fingerprint spectrum according to claim 3, characterized in that: The classification prediction model is constructed based on the quadratic discriminant analysis algorithm.

5. The chemical indicator fingerprint spectrum according to claim 3, characterized in that: The model evaluation indicators include classification accuracy, area under the curve, recall rate, and accuracy.

6. The chemical indicator fingerprint spectrum according to claim 1, characterized in that: It includes 28 characteristic components, among which the components used to identify sewage treatment plants include pazufloxacin mesylate, octyl methoxycinnamate, oleic acid amide, demethyl-formamido-pirimicarb, octocrylene, (4-nonylphenoxy)-acetic acid, propoxur, and benzylsulfuron-methyl; the components used to identify textile printing and dyeing factories include bensulfuron, piperazine, propiconazole, ritalin, olmesartan, butazone, dimethoate, ethoxyquin, TCPP, triadimenol, 4-formamidoantipyrine, sucralose, estriol, neostigmine, climbazole, isopropylamine, N-butylbenzenesulfonamide, sulfadimethoxine, and tapentadol.

7. Application of the chemical indicator fingerprint spectrum as described in any one of claims 1 to 6 in identifying water pollution sources belonging to sewage treatment plants and / or textile printing and dyeing factories.

8. The use according to claim 7, characterized in that: The following steps are involved: (1) Extract information about compounds contained in target water samples; (2) Identify the compound information based on the quadratic discriminant analysis model to determine the source of the target water sample.

9. The use according to claim 8, characterized in that: The step (1) is solid phase extraction, specifically extraction by connecting Oasis HLB, ISOLUTE ENV+, Strata XCW, and Strata XAW solid phase extraction columns in series.

Images