Method and system for identifying farmland runoff pollution in surface water by using microbial fingerprints

Abstract

The invention discloses a method and a system for identifying farmland runoff pollution in surface water by using microbial fingerprints. The method comprises the following steps: collecting microorganism samples of multiple types of water bodies, and obtaining microorganism composition data through 16S rDNA (ribosomal deoxyribonucleic acid) sequencing; based on a sensitivity-specificity screening index, farmland runoff pollution microorganism fingerprints are determined in combination with obligate anaerobic characteristics, and a double-model integrated judgment mechanism is established: an artificial neural network is utilized to process microorganism category data, an XGBoost model is utilized to analyze microorganism relative abundance data, and a classification decision basis of double models is extracted through a control variable method; and performing result judgment based on a logic rule by adopting double models. According to the method, microbial fingerprints and machine learning classification model integration are creatively combined, accurate and simple identification of farmland runoff pollution in surface water can be realized, and the farmland runoff pollution identification accuracy and identification efficiency are improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of environmental pollution source identification, and in particular relates to a method and system for identifying farmland runoff pollution in surface water by using microbial fingerprints. Background Art

[0002] As agricultural production expands, agricultural non-point source pollution has become a significant source of water contamination. Compared to orchards, grasslands, and woodlands, intensively managed farmland has a greater impact on surface water quality. The fertilizers and pesticides used in farmland, through rainfall and other factors, carry pollutants into water bodies, threatening human health and ecological safety. Current pollution source identification methods rely on fingerprints such as stable isotopes, water chemistry indicators, and characteristic compounds. While accurate in some cases, these methods are limited to specific pollutants and cannot distinguish between farmland and other agricultural sources of pollution. Microbial fingerprints, however, offer a promising alternative because they are sensitive to environmental changes and can reflect complex pollution patterns.

[0003] In recent years, research on water pollution identification based on microbial fingerprints has made significant progress. Existing technologies primarily rely on constructing microbial fingerprint databases for specific pollution sources. For example, patent CN117512147A uses the 16S rRNA gene to establish species fingerprints for aquaculture and crop production, and CN117535435A proposes a specific gene fingerprint for rice paddy pollution. However, three major technical bottlenecks remain in this area. First, traditional fingerprint matching methods rely on the detection of microbial fingerprints, which have biological limitations in terms of host specificity and sensitivity, resulting in a high false positive rate. Second, machine learning-based pollution identification focuses on microbial fingerprint abundance data while ignoring the importance of microbial species information. This results in the inability to construct a robust and accurate pollution identification framework due to the indicative role of certain fingerprints in pollution sources. Third, while machine learning models excel at discovering correlations and associations to improve prediction accuracy, they often require cumbersome model construction, optimization, and screening, and lack the ability to perform causal reasoning or statistical interpretation. In particular, neural networks and ensemble models, due to their black-box nature, make decision paths untraceable, making them difficult to meet the requirements of establishing a chain of evidence in environmental law enforcement. Therefore, there is an urgent need to develop an accurate and simple method for identifying pollution sources that integrates multi-dimensional microbial fingerprint features with explainable and transparent machine learning. Summary of the Invention

[0004] In view of the shortcomings of the existing technology, the purpose of the present invention is to provide a method and system for identifying agricultural runoff pollution in surface water using microbial fingerprints, so as to improve the efficiency and accuracy of identifying agricultural runoff pollution.

[0005] According to one aspect of the present application, a method for identifying agricultural runoff pollution in surface water using microbial fingerprints is disclosed, the method comprising:

[0006] Obtain multiple different types of water samples and obtain the pollutant microbial composition data of each water sample based on the 16Sr DNA sequencing method. Each water sample corresponds to a water environment.

[0007] Based on the sensitivity-specificity analysis method and combined with the obligate anaerobic characteristics, multiple farmland runoff pollution-associated microorganisms were screened from the microbial composition data of multiple pollution sources. The farmland runoff pollution-associated microorganisms were used as farmland runoff pollution microbial fingerprints. The farmland runoff pollution microbial fingerprints included f_Desulfuromonadaceae, g_Geobacter, f_AKAU3564_sediment_group, o_Dehalococcoidales, and g_Citrifermentans.

[0008] Based on the farmland runoff pollution microbial category data corresponding to each of the multiple farmland runoff pollution microbial fingerprints and the farmland runoff pollution microbial relative abundance data corresponding to each of the multiple farmland runoff pollution microbial fingerprints, an artificial neural network model and an XGBoost model were constructed respectively;

[0009] Obtaining target data of a target microbial fingerprint in a target surface water sample to be identified, and inputting the target data into the artificial neural network model and the XGBoost model, wherein the target data includes target microbial fingerprint category data corresponding to the target microbial fingerprint and target microbial fingerprint relative abundance data corresponding to the target microorganism;

[0010] Based on the judgment results output by the artificial neural network model and the XGBoost model respectively, logical rules are used to determine whether there is agricultural runoff pollution in the target surface water sample corresponding to the target microbial fingerprint target data.

[0011] In some embodiments, the method further comprises:

[0012] Changing the farmland runoff pollution microbial fingerprint category data input into the artificial neural network model based on a control variable method to obtain a plurality of first output results;

[0013] Based on the plurality of first output results, determining a binarization determination condition satisfying the output of a positive determination result by the artificial neural network model;

[0014] Changing the relative abundance data of the farmland runoff pollution microbial fingerprint input into the XGBoost model based on a control variable method to obtain a plurality of second output results;

[0015] Based on the plurality of second output results, a relative abundance threshold interval satisfying the output of a positive judgment result by the XGBoost model is determined.

[0016] In some embodiments, based on the plurality of first output results, determining the binarization determination condition that satisfies the output of a positive determination result by the artificial neural network model includes:

[0017] Binarizing the farmland runoff pollution microbial fingerprint category data to obtain a binary 0 / 1 matrix, where 0 represents absence and 1 represents presence;

[0018] Based on the binary 0 / 1 matrix, the single farmland runoff pollution microbial fingerprint and / or the combination of multiple farmland runoff pollution microbial fingerprints were analyzed, and the binary judgment conditions were determined to include:

[0019] When the g_Citrifermentans microbial fingerprint is detected, the extraction result is 1, and it is determined that the binary judgment condition is met; or,

[0020] When any two or more microbial fingerprints other than the combination of f_Desulfuromonadaceae and g_Geobacter are detected at the same time, the extraction result is 1, and it is determined that the binarization judgment condition is met.

[0021] In some embodiments, determining, based on the plurality of second output results, a relative abundance threshold interval satisfying the output of a positive judgment result by the XGBoost model comprises:

[0022] The relative abundance data of farmland runoff pollution microbial fingerprints were adjusted according to a precision gradient of 0.0000001%, and the relative abundance threshold intervals were determined to include:

[0023] If only a single microbial fingerprint of agricultural runoff pollution was detected, the relative abundance threshold intervals needed to meet 0.0067339% ≥ f_AKAU3564_sediment_group ≥ 0.0050275%, f_Desulfuromonadaceae ≥ 0.0177225%, 0.0397715% ≥ g_Geobacter ≥ 0.0077625%, 0.0078340% ≥ o_Dehalococcoidales ≥ 0.0031500%, or 0.0105874% ≥ o_Dehalococcoidales ≥ 0.0091936%;

[0024] If f_Desulfuromonadaceae and g_Geobacter were detected at the same time, the relative abundance threshold interval must meet the following requirements: f_Desulfuromonadaceae ≥ 0.0177225% or f_Desulfuromonadaceae < 0.0177225% and 0.0397719% ≥ g_Geobacter ≥ 0.0077645%.

[0025] In some embodiments, the determining whether there is farmland runoff pollution in the target surface water sample corresponding to the target microbial fingerprint target data using logic rules based on the determination results output by the artificial neural network model and the XGBoost model includes:

[0026] Matching the target microorganism fingerprint category data to see whether it satisfies the binarization determination condition;

[0027] When the target microbial fingerprint category data meets the binarization judgment condition and the judgment result output by the artificial neural network model is determined to be a positive judgment result, it is determined that agricultural runoff pollution exists in the target surface water sample corresponding to the target microbial fingerprint category data.

[0028] In some embodiments, the determining whether there is farmland runoff pollution in the target surface water sample corresponding to the target microbial fingerprint target data using logic rules based on the determination results output by the artificial neural network model and the XGBoost model includes:

[0029] When the determination result output by the artificial neural network model is a negative determination result;

[0030] Obtaining relative abundance data of target microbial fingerprints in target surface water samples;

[0031] whether the relative abundance data matching the target microbial fingerprint meets the relative abundance threshold interval;

[0032] When the relative abundance data of the target microbial fingerprint meets the relative abundance threshold interval, the judgment result output by the XGBoost model is determined to be a positive judgment result, and it is determined that agricultural runoff pollution exists in the target surface water sample corresponding to the relative abundance data of the target microbial fingerprint.

[0033] In some embodiments, the determining whether there is farmland runoff pollution in the target surface water sample corresponding to the target microbial fingerprint target data using logic rules based on the determination results output by the artificial neural network model and the XGBoost model includes:

[0034] Matching the target microorganism fingerprint category data to see whether it satisfies the binarization determination condition;

[0035] Whether the relative abundance data of the target microbial fingerprint in the target surface water sample meets the relative abundance threshold interval;

[0036] When the target microbial fingerprint category data satisfies the binarization determination condition and / or the relative abundance data of the target microbial fingerprint satisfies the relative abundance threshold interval, it is determined whether there is farmland pollution in the target surface water sample corresponding to the target microbial fingerprint target data.

[0037] According to another aspect of the present application, a system for identifying agricultural runoff pollution in surface water using microbial fingerprints is also disclosed, the system comprising:

[0038] The water sample acquisition module is used to obtain multiple different types of water samples and obtain the microbial composition data of each water sample based on the 16S rDNA sequencing method. Each water sample corresponds to a water environment;

[0039] A farmland runoff pollution microbial fingerprint determination module is used to screen multiple farmland runoff pollution-associated microorganisms from multiple pollution source microbial composition data based on a sensitivity-specificity analysis method and in combination with obligate anaerobic characteristics, and use the farmland runoff pollution-associated microorganisms as farmland runoff pollution microbial fingerprints, wherein the farmland runoff pollution microbial fingerprints include f_Desulfuromonadaceae, g_Geobacter, f_AKAU3564_sediment_group, o_Dehalococcoidales, and g_Citrifermentans;

[0040] A model building module is used to build an artificial neural network model and an XGBoost model based on the farmland runoff pollution microbial fingerprint category data corresponding to each of the multiple farmland runoff pollution microbial fingerprints and the farmland runoff pollution microbial fingerprint relative abundance data corresponding to each of the multiple farmland runoff pollution microbial fingerprints;

[0041] a data acquisition module, configured to acquire target data of a target microbial fingerprint in a target surface water sample to be identified, and input the target data into the artificial neural network model and the XGBoost model, wherein the target data includes target microbial fingerprint category data corresponding to the target microbial fingerprint and target microbial fingerprint relative abundance data corresponding to the target microbial fingerprint;

[0042] The farmland runoff determination module is used to determine whether farmland runoff pollution exists in the target surface water sample corresponding to the target microbial fingerprint target data based on the judgment results output by the artificial neural network model and the XGBoost model, using logical rules.

[0043] The present invention includes but is not limited to the following beneficial effects: (1) The present invention utilizes 16S rDNA sequencing was used to determine the microbial composition of samples from various pollution sources. Combined with sensitivity-specificity analysis and obligate anaerobic characteristics, multiple farmland runoff pollution-associated microorganisms were screened out. Based on the farmland runoff pollution-associated microorganisms, a microbial fingerprint for identifying farmland runoff pollution was constructed, filling the gap in the field of farmland runoff pollution identification in surface water due to the lack of specific microbial indicators; (2) The present invention provides a simple model-free method for identifying farmland runoff pollution in surface water using microbial fingerprints, breaking through the dependence of pollution source identification on machine learning algorithm modeling, replacing complex data analysis with explicit judgment rules, simplifying the pollution source identification process, and improving the operability of the pollution identification method; (3) The present invention proposes a two-dimensional fusion criterion for microbial fingerprints to improve the robustness of pollution identification. By jointly analyzing the category and relative abundance threshold of microbial fingerprints, it effectively overcomes the limitation that a single data model is difficult to fully capture the complex information of the microbiome; (4) The technology of the present invention is highly scalable. The constructed pollution source identification framework (microbial fingerprint screening + two-dimensional criterion) can be migrated to other pollution source identification scenarios such as domestic sewage and industrial wastewater, and has high application value in environmental pollution source analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for describing the embodiments or the prior art.

[0045] Figure 1 This is a flow chart of a method for identifying agricultural runoff pollution in surface water using microbial fingerprints according to an embodiment of the present application;

[0046] Figure 2 This is another flow chart of a method for identifying agricultural runoff pollution in surface water using microbial fingerprints according to an embodiment of the present application;

[0047] Figure 3 This is another flow chart of a method for identifying agricultural runoff pollution in surface water using microbial fingerprints according to an embodiment of the present application;

[0048] Figure 4 This is another flow chart of a method for identifying agricultural runoff pollution in surface water using microbial fingerprints according to an embodiment of the present application;

[0049] Figure 5 This is another flow chart of a method for identifying agricultural runoff pollution in surface water using microbial fingerprints according to an embodiment of the present application;

[0050] Figure 6 This is a structural block diagram of a system for identifying agricultural runoff pollution in surface water using microbial fingerprints according to an embodiment of the present application;

[0051] Figure 7 This is an NMDS dimensionality reduction grouping diagram of various pollution source categories based on the Bray-Curtis distance in an embodiment of the present invention. The pollution source categories on the left can be significantly divided into farmland runoff pollution and domestic sewage, the pollution source categories in the middle can be significantly divided into farmland runoff pollution and aquaculture tailwater, and the pollution source categories on the right can be significantly divided into farmland runoff pollution and garden runoff pollution.

[0052] Figure 8 The balanced accuracy and F1score of the classification data model based on microbial fingerprints in an embodiment of the present invention are shown on the left, where the balanced accuracy is shown on the right, and the F1score is shown on the right.

[0053] Figure 9 The balanced accuracy and F1score of the relative abundance data model based on microbial fingerprints in the embodiment of the present invention are shown on the left, where the balanced accuracy is shown on the right, and the F1score is shown on the right.

[0054] Figure 10 This is a performance comparison of the model ensemble, single model, and traditional methods of the embodiment of the present invention on the external test set 2;

[0055] Figure 11 This is the detection of microbial fingerprints of agricultural runoff pollution in surface water samples of an embodiment of the present invention, with seawater samples on the left and river water samples on the right. Des, Geo, AKA, Deh and Cit are the abbreviations of f_Desulfuromonadaceae, g_Geobacter, f_AKAU3564_sediment_group, o_Dehalococcoidales and g_Citrifermentans, respectively. DETAILED DESCRIPTION

[0056] The embodiment of the present invention provides a method for identifying farmland runoff pollution in surface water by using microbial fingerprints, the method comprising: obtaining a plurality of water samples of different types, and obtaining microbial composition data of each water sample based on a 16S rDNA sequencing method, wherein each water sample corresponds to a water environment; based on a sensitivity-specificity analysis method, combined with obligate anaerobic characteristics, screening a plurality of farmland runoff pollution-associated microorganisms from the microbial composition data of a plurality of pollution sources, and using the farmland runoff pollution-associated microorganisms as farmland pollution microbial fingerprints, wherein the farmland runoff pollution microbial fingerprints include f_Desulfuromonadaceae, g_Geobacter, f_AKAU3564_sediment_group, o_Dehalococcoidales and g_Citrifermentans; based on the farmland runoff pollution microbial fingerprints corresponding to each of the plurality of farmland runoff pollution microbial fingerprints, An artificial neural network model and an XGBoost model are constructed based on the pollution microbial fingerprint category data and the relative abundance data of the farmland runoff pollution microbial fingerprints corresponding to each of the multiple farmland runoff pollution microbial fingerprints. The target data of the target microbial fingerprint in the target surface water sample to be identified is obtained, and the target data is input into the artificial neural network model and the XGBoost model, wherein the target data includes the target microbial fingerprint category data corresponding to the target microbial fingerprint and the target microbial fingerprint relative abundance data corresponding to the target microbial fingerprint. Based on the judgment results output by the artificial neural network model and the XGBoost model, a logical rule is used to determine whether farmland runoff pollution exists in the target surface water sample corresponding to the target microbial fingerprint target data. This application improves the efficiency and accuracy of identifying farmland runoff pollution.

[0057] The terms "first," "second," "third," "fourth," and the like (if any) in the description and claims of the present invention and in the accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a particular order or precedence. It should be understood that the terms used in this manner are interchangeable where appropriate so that the embodiments described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "including" or "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product, or apparatus that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units that are not explicitly listed or that are inherent to these processes, methods, products, or apparatus.

[0058] For ease of understanding, the specific process of the embodiment of the present invention is described below. Specifically, Figure 1 A flow chart of a method for identifying agricultural runoff pollution in surface water using microbial fingerprints, such as Figure 1 As shown, the following steps are included:

[0059] S100. Obtain multiple water samples of different types, and obtain microbial composition data of each water sample based on the 16S rDNA sequencing method.

[0060] Among them, each water sample corresponds to a water environment.

[0061] Specifically, in one example, this step can be implemented based on the following steps:

[0062] (1) Water sample collection: 158 farmland runoff samples, 37 garden runoff samples, 17 domestic sewage samples, 174 aquaculture tailwater samples, and 186 surface water samples were collected in 1 L volumes, stored at 4°C, transported back to the laboratory, and processed within 24 hours.

[0063] (2) Water sample pretreatment: Filter the water sample through a 0.45 μm glass fiber filter membrane. After filtration, remove the filter membrane and place it in a 2 mL centrifuge tube and store it in a -80°C refrigerator. 16s rDNA high-throughput sequencing experiments were completed on the Illumina platform of Hangzhou Lianchuan Biotechnology Co., Ltd.

[0064] (3) Extraction of total microbial DNA: The CTAB method was used to extract total DNA from microbial groups from samples of various sources. The quality of DNA extraction was tested by agarose gel electrophoresis, and the DNA was quantified using an ultraviolet spectrophotometer.

[0065] (4) PCR amplification and sequencing: The V3-V4 region of the 16s rDNA gene was amplified using the DNA extracted from each sample as a template. The sequencing primer sequences were:

[0066] Forward primer: 341F (5′-CCTACGGGNGGCWGCAG-3′);

[0067] Rear primer: 805R (5'-GACTACHVGGGTATCTAATCC-3').

[0068] The specific reaction system and conditions are shown in Table 1 and Table 2:

[0069] Table 1 PCR reaction system

[0070] PCR reaction components PCR reaction volume PhusionHotstartflex2XMasterMix 12.5μL ForwardPrimer 2.5 μL ReversePrimer 2.5 μL Template DNA 20ng <![CDATA[ddH2O]]> 25 μL

[0071] Table 2 PCR reaction conditions

[0072] PCR reaction temperature PCR reaction time Number of cycles 98℃ 30s 98℃ 10s 54℃ 30s 35 cycles 72℃ 45s 72℃ 10min

[0073]

[0074] The purified and quality-checked PCR products were mixed in the same proportion and sequenced on the Illumina NovaSeq platform.

[0075] (5) ASV clustering and species annotation: The obtained microbial gene sequences were clustered and abundance determined based on 100% similarity, and the obtained abundance results were normalized. To obtain the species classification information corresponding to each ASV, species classification was performed using the SILVA and NT-16s databases. To ensure the accuracy of the classification annotation results, the annotation confidence level was set to greater than 0.7. Based on the ASV annotation results and the ASV abundance table for each sample, the relative abundance of species at the kingdom, phylum, class, order, family, genus, and species levels was obtained.

[0076] S102. Based on the sensitivity-specificity analysis method and combined with the obligate anaerobic characteristics, multiple farmland runoff pollution-associated microorganisms are screened out from the microbial composition data of multiple pollution sources, and the farmland runoff pollution-associated microorganisms are used as the farmland runoff pollution microbial fingerprints.

[0077] In one example, the agricultural runoff pollution microbial fingerprint includes f_Desulfuromonadaceae, g_Geobacter, f_AKAU3564_sediment_group, o_Dehalococcoidales, and g_Citrifermentans.

[0078] Specifically, in combination with the above steps (1) to (5), proceed to step (6) to determine the microbial fingerprint.

[0079] (6) Based on the count values ​​of ASVs, the Bray-Curtis distance was calculated using the vegdist function of R language, and the dimensionality reduction grouping of pollution source samples was performed based on the Bray-Curtis distance using non-metric multidimensional scaling (NMDS). The results are as follows: Figure 7 As shown in Table 3, the results showed that the pollution source samples could be significantly divided into farmland runoff, aquaculture tailwater, garden runoff, and domestic sewage. Ten specific microorganisms for farmland runoff were identified through source sensitivity (≥0.5) and source specificity (≥0.8), as shown in Table 3. Combining the obligate anaerobic characteristics, five specific anaerobic microorganisms were identified as the microbial fingerprints for farmland runoff pollution: f_Desulfuromonadaceae, g_Geobacter, f_AKAU3564_sediment_group, o_Dehalococcoidales, and g_Citrifermentans.

[0080] Table 3 Sensitivity and specificity of specific microorganisms for farmland pollution

[0081]

[0082] S104. Based on the farmland runoff pollution microbial fingerprint category data corresponding to each of the multiple farmland runoff pollution microbial fingerprints and the farmland runoff pollution microbial fingerprint relative abundance data corresponding to each of the multiple farmland runoff pollution microbial fingerprints, an artificial neural network model and an XGBoost model are constructed respectively.

[0083] Specifically, the model construction can refer to the following steps:

[0084] (a) Configuration of simulated pollution sinks: Based on the target region's hydrological parameters (average annual rainfall, arable land irrigation volume, effective agricultural irrigation water utilization coefficient, surface water resources, and cultivated land area), the proportion of farmland runoff pollution was calculated to be 0% to 20% of the total surface water resources. Farmland runoff pollution was mixed with other pollution sources (garden runoff, domestic sewage, and aquaculture tailwater) in a gradient of proportions (0%, 5%, 10%, and 20%) to create 104 simulated pollution sinks. When constructing the composite simulation system, surface water with no microbial fingerprints of farmland runoff pollution was selected as the background water sample. This was mixed with water samples from various pollution sources in a preset proportion (total volume of 1 L) to construct composite simulation systems containing single, binary, ternary, and quaternary pollution sources.

[0085] (b) Data Preparation: The microbial fingerprint data of all source samples (n = 386) served as the base dataset. The category and relative abundance data of the microbial fingerprints served as the model input features, and the pollution source type served as the output variable. In the category data, the detection of the microbial fingerprint was recorded as 1, and the non-detection was recorded as 0. The output variable was also binary: 0 indicated the absence of agricultural runoff pollution, and 1 indicated the presence of agricultural runoff pollution. The base dataset was divided into a training set (n = 272, containing 108 positive samples and 164 negative samples) and a test set 1 (n = 114, containing 45 positive samples and 69 negative samples) in a 7:3 ratio. The microbial fingerprint data of agricultural runoff pollution from 104 simulated pollution sinks served as an external test set 2 (n = 104, containing 79 positive samples and 25 negative samples) to further evaluate the model's generalization ability, universality, and stability across different data.

[0086] (c) Model Construction: Seven classification models were constructed using the R language, including logistic regression (LR), support vector machine (SVM), naive Bayes (NB), artificial neural network (ANN), random forest (RF), XGBoost, and K-nearest neighbor (KNN). To assess the robustness and randomness of the models, 15 random seeds were created, generating 15 different dataset partitions for independent computation. Before training each model, hyperparameter optimization was performed using a grid search method and 10-fold cross-validation. In this example, an artificial neural network (ANN) and an XGBoost model were used. ANN models form a complex network structure with a large number of interconnected neurons. During training, two types of microbial fingerprint data were input, and the model continuously adjusted the connection weights between neurons. In this process, the ANN model learned the inherent relationship between microbial fingerprint data and agricultural runoff pollution, such as the correlation pattern between specific microbial assemblages and their abundance levels and agricultural runoff pollution. XGBoost Model: As a gradient boosting tree algorithm, XGBoost iteratively constructs a series of decision trees. Each iteration focuses on samples that were incorrectly predicted by the previous model to continuously optimize the model. During training, it builds a decision tree based on microbial fingerprint data and mines the features and patterns in the data to achieve accurate prediction of agricultural runoff pollution.

[0087] (d) Model evaluation: The classification model was evaluated using the following four indicators: Balanced Accuracy, Sensitivity, Specificity, Precision, and F1score. In order to determine the best contamination identification model and reduce the redundancy of evaluation indicators, balanced accuracy and F1score were used as the final screening indicators. The results are shown in the figure below. Figure 8 and Figure 9 The results showed that the ANN model performed best based on the microbial fingerprint classification data, with a balanced accuracy and F1 score of 0.8133±0.0006 and 0.7891±0.0062, respectively. The XGBoost model performed best based on the microbial fingerprint relative abundance data, with a balanced accuracy and F1 score of 0.8261±0.0029 and 0.8105±0.0057, respectively.

[0088] (e) Model integration: To ensure the representativeness of the dataset, we used the basic dataset (n=386) as the training set and the external test set 2 (n=104) as the test set. After hyperparameter optimization, the parameters were set as follows: for the ANN model, size, decay, maxit and abstol were set to 1, 0.001, 500 and 0.001, respectively; for the XGBoost model, max_depth, eta, nrounds, gamma, subsample and colsample_bytree were set to 6, 0.01, 172, 0, 1 and 1, respectively. ANN and XGBoost were integrated using the logical "or" rule, and if the output of any model was 1, it was judged to be agricultural runoff pollution. We compared the performance of the model ensemble with that of a single model and the traditional method (where the detection of microbial fingerprints was considered to be the presence of agricultural runoff pollution) on the external test set 2, as shown in Figure 2. Figure 10 The results show that the performance of the model ensemble is significantly improved, with the balanced accuracy and F1 score reaching 0.8661 and 0.8522, respectively. Compared with the highest value of a single model, these increases are 2.24% and 2.97%, respectively, and compared with traditional methods, they are 17.97% and 14.28%, respectively.

[0089] S106. Obtain target data of target microbial fingerprints in the target surface water sample to be identified, and input the target data into the artificial neural network model and the XGBoost model.

[0090] The target data includes target microbial fingerprint category data corresponding to the target microbial fingerprint and target microbial fingerprint relative abundance data corresponding to the target microbial fingerprint.

[0091] S108. Based on the judgment results output by the artificial neural network model and the XGBoost model, logical rules are used to determine whether there is agricultural runoff pollution in the target surface water sample corresponding to the target microbial fingerprint target data.

[0092] It can be understood that this embodiment uses 16S rDNA sequencing to determine the microbial composition of samples from various types of pollution sources, combines sensitivity-specificity analysis and obligate anaerobic characteristics to screen out multiple farmland runoff pollution-associated microorganisms, and constructs a microbial fingerprint for identifying farmland runoff pollution based on farmland runoff pollution-associated microorganisms, filling the gap in the field of farmland runoff pollution identification in surface water due to the lack of specific microbial indicators; and this simple model-free method for identifying farmland runoff pollution in surface water using microbial fingerprints breaks through the dependence of pollution source identification on machine learning algorithm modeling, replaces complex data analysis with explicit judgment rules, simplifies the pollution source identification process, and improves the operability of the pollution identification method; proposes a two-dimensional fusion criterion for microbial fingerprints to improve the robustness of pollution identification, and effectively overcomes the limitation that a single data model is difficult to fully capture the complex information of the microbiome by jointly analyzing the category and relative abundance threshold of the microbial fingerprint.

[0093] Further, Figure 2 This is another flow chart of the method for identifying agricultural runoff pollution in surface water using microbial fingerprints according to an embodiment of the present application. Figure 2 As shown, the following steps are included:

[0094] S200: Changing the farmland runoff pollution microbial fingerprint category data input into the artificial neural network model based on the control variable method to obtain multiple first output results.

[0095] S202: Based on the multiple first output results, determine a binarization determination condition that satisfies the output of a positive determination result by the artificial neural network model.

[0096] S204. Based on the control variable method, the relative abundance data of the farmland runoff pollution microbial fingerprint input into the XGBoost model is changed to obtain multiple second output results.

[0097] S206. Based on the multiple second output results, determine a relative abundance threshold interval that satisfies the XGBoost model outputting a positive judgment result.

[0098] Specifically, for the artificial neural network model (ANN model), the control variable method was used. While holding all other variables constant, only one category of farmland runoff contamination microbial fingerprint data was changed (set to present or absent), and the changes in the model output were observed. Through multiple experiments, it was determined that under which combinations of microbial fingerprint categories, the ANN model would output a positive judgment (i.e., the presence of farmland runoff contamination). This determined the binary judgment conditions that satisfied the ANN model's output of a positive judgment result. "Binary" here means that the microbial fingerprint category has only two states: "present" (which can be considered 1) and "absent" (which can be considered 0).

[0099] In one example, suppose five microbial fingerprints of farmland pollution are identified: A, B, C, D, and E. Experiments using the controlled variable method revealed that when microbial fingerprints A, B, and C are present, and D and E are absent, the ANN model outputs a positive judgment, indicating the presence of farmland pollution in the sample. Therefore, "A, B, and C present, and D and E absent" is a binary judgment condition. In practice, there are multiple such combinations of conditions, which together form the basis for the ANN model to determine farmland runoff pollution. These conditions provide clear rules for the subsequent use of the ANN model to identify farmland runoff pollution in surface water samples.

[0100] For the XGBoost model, we also used the control variable method, fixing the category data of the microbial fingerprint and changing its relative abundance data. We gradually adjusted the relative abundance values ​​and observed the model output to determine the ranges in which the relative abundance of the microbial fingerprint would result in the XGBoost model outputting a positive judgment, thereby obtaining the relative abundance threshold range.

[0101] Furthermore, in one example, for the ANN model, whether it is a single farmland runoff pollution microbial fingerprint or a combination of multiple farmland runoff pollution microbial fingerprints, the data of the target farmland runoff pollution microbial fingerprint is set to 1 to indicate its detection, and the data of other fingerprints are uniformly set to 0 to indicate non-detection. Then, the impact of each fingerprint and fingerprint combination on the model prediction was explored in turn. In one example, in order to address the limitations of the ANN model, the XGBoost model was set as follows: when a specific single fingerprint or fingerprint combination is detected but not identified as farmland pollution by the ANN model, the relative abundance range of the fingerprint identified by the XGBoost model as farmland runoff pollution is determined by changing the relative abundance of the fingerprint. The relative abundance range can be set from 0.002% to 100%. In order to minimize errors, the relative abundance data is accurate to seven decimal places, and the data of two adjacent samples are increased by 0.0000001% in sequence. The results showed that for the ANN model, if g_Citrifermentans or any two or more microbial fingerprints other than the combination of f_Desulfuromonadaceae and g_Geobacter were detected at the same time, it would be identified as the presence of agricultural runoff pollution. For the XGBoost model, if only a single microbial fingerprint was detected, it would be identified as the presence of agricultural runoff pollution if any of the following conditions were met: 0.0067339% ≥ f_AKAU3564_sediment_group ≥ 0.0050275%, f_Desulfuromonadaceae ≥ 0.0177225%, 0.0397715% ≥ g_Geobacter ≥ 0.0077625%, 0.0078340% ≥ o_Dehalococcoidales ≥ 0.0031500% or 0.0105874% ≥ o_Dehalococcoidales ≥ 0.0091936%; if f_Desulfuromonadaceae and g_Geobacter are detected at the same time, it will be identified as agricultural runoff pollution if any of the following conditions is met: f_Desulfuromonadaceae ≥ 0.0177225% or f_Desulfuromonadaceae < 0.0177225% and 0.0397719% ≥ g_Geobacter ≥ 0.0077645%.

[0102] Through clear binarization and specific judgment criteria, the ANN model can quickly analyze and judge samples without complex calculations, improving the efficiency of identifying farmland runoff pollution. In actual environmental monitoring, preliminary judgment results can be quickly provided, buying time for subsequent processing. Judgment criteria are constructed from two dimensions: single fingerprints and multiple fingerprint combinations. This not only focuses on the indicative role of key single microorganisms, but also considers the overall characteristics of the microbial community. This comprehensive multi-angle judgment reduces misjudgments and missed judgments, thereby improving the model's accuracy and reliability in identifying farmland pollution.

[0103] In one example, Figure 3 This is another flow chart of the method for identifying agricultural runoff pollution in surface water using microbial fingerprints according to an embodiment of the present application. Figure 3 As shown, the following steps are included:

[0104] S300: Match the target microorganism fingerprint category data to see whether it meets the binarization determination conditions.

[0105] S302: When the target microbial fingerprint category data satisfies the binarization determination condition and the determination result output by the artificial neural network model is determined to be a positive determination result, it is determined that farmland runoff pollution exists in the target surface water sample corresponding to the target microbial fingerprint category data.

[0106] In one example, Figure 4 This is another flow chart of the method for identifying agricultural runoff pollution in surface water using microbial fingerprints according to an embodiment of the present application. Figure 4 As shown, the following steps are included:

[0107] S400: When the determination result output by the artificial neural network model is a negative determination result, relative abundance data of the target microbial fingerprint in the target surface water sample is obtained.

[0108] S402: Whether the relative abundance data of the target microbial fingerprint meets the relative abundance threshold range.

[0109] S404. When the relative abundance data of the target microbial fingerprint meets the relative abundance threshold interval, the judgment result output by the XGBoost model is determined to be a positive judgment result, and it is determined that agricultural runoff pollution exists in the target surface water sample corresponding to the relative abundance data of the target microbial fingerprint.

[0110] In one example, Figure 5 This is another flow chart of the method for identifying agricultural runoff pollution in surface water using microbial fingerprints according to an embodiment of the present application. Figure 5 As shown, the following steps are included:

[0111] S500: Match the target microorganism fingerprint category data to see if it meets the binarization criteria.

[0112] S502: Match the target microbial fingerprint relative abundance data in the target surface water sample to see if it meets the relative abundance threshold range.

[0113] S504. When the target microbial fingerprint category data meets the binarization judgment condition and / or the target microbial fingerprint relative abundance data meets the relative abundance threshold interval, a logical rule is used to determine whether there is agricultural runoff pollution in the target surface water sample corresponding to the target microbial fingerprint target data.

[0114] Based on this example method, in one example, the target data of the target microbial fingerprint in the target surface water sample to be identified is as follows: Figure 11 As shown, the results showed that 47 of the 186 surface water samples collected detected microbial fingerprints indicating agricultural runoff pollution.

[0115] The target data of the target microbial fingerprints in the target surface water samples to be identified were compared with the judgment conditions, and the results are shown in Table 4. The results showed that 32 of the 47 surface water samples with detected microbial fingerprints were identified as having agricultural runoff pollution.

[0116] Table 4 Identification results of agricultural runoff pollution in surface water. 1 represents the presence of agricultural runoff pollution, and 0 represents the absence of agricultural runoff pollution.

[0117]

[0118] Further, Figure 6 The structural diagram of the system for identifying agricultural runoff pollution in surface water using microbial fingerprints is shown in Figure 2. Figure 6 As shown, the system includes:

[0119] The water sample acquisition module is used to obtain multiple different types of water samples and obtain the microbial composition data of each water sample based on the 16S rDNA sequencing method. Each water sample corresponds to a water environment;

[0120] The farmland runoff pollution microbial fingerprint determination module is used to screen multiple farmland runoff pollution-associated microorganisms from the microbial composition data of multiple pollution sources based on the sensitivity-specificity analysis method and combined with the obligate anaerobic characteristics. The farmland runoff pollution-associated microorganisms are used as the farmland runoff pollution microbial fingerprint. The farmland runoff pollution microbial fingerprint includes f_Desulfuromonadaceae, g_Geobacter, f_AKAU3564_sediment_group, o_Dehalococcoidales, and g_Citrifermentans;

[0121] A model building module is used to build an artificial neural network model and an XGBoost model based on the farmland runoff pollution microbial fingerprint category data corresponding to each of the multiple farmland runoff pollution microbial fingerprints and the farmland runoff pollution microbial fingerprint relative abundance data corresponding to each of the multiple farmland runoff pollution microbial fingerprints;

[0122] a data acquisition module, configured to acquire target data of a target microbial fingerprint in a target surface water sample to be identified, and input the target data into an artificial neural network model and an XGBoost model, wherein the target data includes target microbial fingerprint category data corresponding to the target microbial fingerprint and relative abundance data of the target microbial fingerprint corresponding to the target microbial fingerprint;

[0123] The farmland runoff pollution determination module is used to determine whether farmland runoff pollution exists in the target surface water sample corresponding to the target microbial fingerprint target data based on the judgment results output by the artificial neural network model and the XGBoost model, using logical rules.

[0124] The application introduction of the relevant modules of the system in this example can refer to the relevant introduction of the above method principles, which will not be repeated here.

[0125] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A method for identifying agricultural runoff pollution in surface water using microbial fingerprints, characterized in that: The method comprises: Obtain multiple different types of water samples and obtain the pollution source microbial composition data of each water sample based on the 16S rDNA sequencing method. Each water sample corresponds to a water environment. Based on the sensitivity-specificity analysis method and combined with the obligate anaerobic characteristics, multiple farmland runoff pollution-associated microorganisms were screened from the microbial composition data of multiple pollution sources. The farmland runoff pollution-associated microorganisms were used as farmland runoff pollution microbial fingerprints. The farmland runoff pollution microbial fingerprints included f_Desulfuromonadaceae, g_Geobacter, f_AKAU3564_sediment_group, o_Dehalococcoidales, and g_Citrifermentans. Based on the farmland runoff pollution microbial fingerprint category data corresponding to each of the multiple farmland runoff pollution microbial fingerprints and the farmland runoff pollution microbial fingerprint relative abundance data corresponding to each of the multiple farmland runoff pollution microbial fingerprints, an artificial neural network model and an XGBoost model were constructed respectively; Obtaining target data of a target microbial fingerprint in a target surface water sample to be identified, and inputting the target data into the artificial neural network model and the XGBoost model, wherein the target data includes target microbial fingerprint category data corresponding to the target microbial fingerprint and target microbial fingerprint relative abundance data corresponding to the target microbial fingerprint; Based on the judgment results output by the artificial neural network model and the XGBoost model respectively, logical rules are used to determine whether there is agricultural runoff pollution in the target surface water sample corresponding to the target microbial fingerprint target data.

2. The method for identifying agricultural runoff pollution in surface water using microbial fingerprints according to claim 1, characterized in that: The method further comprises: Changing the farmland runoff pollution microbial fingerprint category data input into the artificial neural network model based on a control variable method to obtain a plurality of first output results; Based on the plurality of first output results, determining a binarization determination condition satisfying the output of a positive determination result by the artificial neural network model; Changing the relative abundance data of the farmland runoff pollution microbial fingerprint input into the XGBoost model based on a control variable method to obtain a plurality of second output results; Based on the plurality of second output results, a relative abundance threshold interval satisfying the output of a positive judgment result by the XGBoost model is determined.

3. The method for identifying agricultural runoff pollution in surface water using microbial fingerprints according to claim 2, characterized in that: Determining, based on the plurality of first output results, a binarization determination condition satisfying the output of a positive determination result by the artificial neural network model includes: Binarizing the farmland runoff pollution microbial fingerprint category data to obtain a binary 0 / 1 matrix, where 0 represents absence and 1 represents presence; Based on the binary 0 / 1 matrix, the single farmland runoff pollution microbial fingerprint and / or the combination of multiple farmland runoff pollution microbial fingerprints were analyzed, and the binary judgment conditions were determined to include: When the g_Citrifermentans microbial fingerprint is detected, the extraction result is 1, and it is determined that the binary judgment condition is met; or, When any two or more microbial fingerprints other than the combination of f_Desulfuromonadaceae and g_Geobacter are detected at the same time, the extraction result is 1, and it is determined that the binarization judgment condition is met.

4. The method for identifying agricultural runoff pollution in surface water using microbial fingerprints according to claim 2, characterized in that: Determining the relative abundance threshold interval that satisfies the output of a positive judgment result by the XGBoost model based on the plurality of second output results includes: The relative abundance data of farmland runoff pollution microbial fingerprints were adjusted according to a precision gradient of 0.0000001%, and the relative abundance threshold intervals were determined to include: If only a single microbial fingerprint of agricultural runoff pollution was detected, the relative abundance threshold intervals needed to meet 0.0067339% ≥ f_AKAU3564_sediment_group ≥ 0.0050275%, f_Desulfuromonadaceae ≥ 0.0177225%, 0.0397715% ≥ g_Geobacter ≥ 0.0077625%, 0.0078340% ≥ o_Dehalococcoidales ≥ 0.0031500%, or 0.0105874% ≥ o_Dehalococcoidales ≥ 0.0091936%; If f_Desulfuromonadaceae and g_Geobacter were detected at the same time, the relative abundance threshold interval must meet the following requirements: f_Desulfuromonadaceae ≥ 0.0177225% or f_Desulfuromonadaceae < 0.0177225% and 0.0397719% ≥ g_Geobacter ≥ 0.0077645%.

5. The method for identifying agricultural runoff pollution in surface water using microbial fingerprints according to claim 1, characterized in that: The determination based on the judgment results outputted by the artificial neural network model and the XGBoost model, using logic rules to determine whether there is farmland runoff pollution in the target surface water sample corresponding to the target microbial fingerprint category data, includes: Matching the target microorganism fingerprint category data to see whether it satisfies the binarization determination condition; When the target microbial fingerprint category data meets the binarization judgment condition and the judgment result output by the artificial neural network model is determined to be a positive judgment result, it is determined that agricultural runoff pollution exists in the target surface water sample corresponding to the target microbial fingerprint category data.

6. The method for identifying agricultural runoff pollution in surface water using microbial fingerprints according to claim 1, characterized in that: The determination based on the judgment results outputted by the artificial neural network model and the XGBoost model, using logic rules to determine whether there is farmland runoff pollution in the target surface water sample corresponding to the target microbial fingerprint target data, includes: When the determination result output by the artificial neural network model is a negative determination result, obtaining relative abundance data of the target microbial fingerprint in the target surface water sample; whether the relative abundance data matching the target microbial fingerprint meets the relative abundance threshold interval; When the relative abundance data of the target microbial fingerprint meets the relative abundance threshold interval, the judgment result output by the XGBoost model is determined to be a positive judgment result, and it is determined that agricultural runoff pollution exists in the target surface water sample corresponding to the relative abundance data of the target microbial fingerprint.

7. The method for identifying agricultural runoff pollution in surface water using microbial fingerprints according to claim 1, characterized in that: The determination based on the judgment results outputted by the artificial neural network model and the XGBoost model, using logic rules to determine whether there is farmland runoff pollution in the target surface water sample corresponding to the target microbial fingerprint target data, includes: Matching the target microorganism fingerprint category data to see whether it satisfies the binarization determination condition; Whether the relative abundance data of the target microbial fingerprint in the target surface water sample meets the relative abundance threshold interval; When the target microbial fingerprint category data meets the binarization judgment condition and / or the relative abundance data of the target microbial fingerprint meets the relative abundance threshold interval, it is determined whether there is agricultural runoff pollution in the target surface water sample corresponding to the target microbial fingerprint target data.

8. A system for identifying agricultural runoff pollution in surface water using microbial fingerprints, characterized in that: The system comprises: The water sample acquisition module is used to obtain multiple different types of water samples and obtain the microbial composition data of each water sample based on the 16S rDNA sequencing method. Each water sample corresponds to a water environment; A farmland runoff pollution microbial fingerprint determination module is used to screen multiple farmland runoff pollution-associated microorganisms from multiple pollution source microbial composition data based on a sensitivity-specificity analysis method and in combination with obligate anaerobic characteristics, and use the farmland runoff pollution-associated microorganisms as farmland runoff pollution microbial fingerprints, wherein the farmland runoff pollution microbial fingerprints include f_Desulfuromonadaceae, g_Geobacter, f_AKAU3564_sediment_group, o_Dehalococcoidales, and g_Citrifermentans; A model building module is used to build an artificial neural network model and an XGBoost model based on the farmland runoff pollution microbial fingerprint category data corresponding to each of the multiple farmland runoff pollution microbial fingerprints and the farmland runoff pollution relative abundance data corresponding to each of the multiple farmland runoff pollution microbial fingerprints; a data acquisition module, configured to acquire target data of a target microbial fingerprint in a target surface water sample to be identified, and input the target data into the artificial neural network model and the XGBoost model, wherein the target data includes target microbial fingerprint category data corresponding to the target microorganism and target microbial relative abundance data corresponding to the target microorganism; The farmland runoff pollution determination module is used to determine whether farmland runoff pollution exists in the target surface water sample corresponding to the target microbial fingerprint target data based on the judgment results output by the artificial neural network model and the XGBoost model, using logical rules.

Images