Prediction model construction method for predicting cell density of drug-resistant pathogenic bacteria based on physicochemical indexes of water quality and application of prediction model construction method
Through the construction of a prediction model based on water quality physical and chemical indicators, the gradient lift algorithm is used to predict the cell density of drug-resistant pathogenic bacteria in rivers, which solves the problem of difficult to monitor and identify drug-resistant pathogenic bacteria in the existing technology in real time, and achieves drug-resistant resistance Effective monitoring of pathogenic bacteria risks and environmental factor analysis promote their risk reduction and health protection.
Patent Information
- Application Number
- CN202410914823.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-09
- Publication Date
- 2024-12-03
AI Technical Summary
The existing technology cannot effectively monitor and quickly identify the risks of drug-resistant pathogenic bacteria in rivers in real time, and its environmental drivers and their impact mechanisms are unclear, resulting in the inability to effectively reduce the risk of drug-resistant pathogenic bacteria from the control of environmental factors. .
Using the prediction model construction method based on water quality physical and chemical indexes, a regression model is established through the gradient elevator algorithm, and using the physical and chemical index data in water quality monitoring (such as carbamazepine, ibuprofen, temperature, redox potential, etc.) as explanatory variables to predict Cell density of drug-resistant pathogenic bacteria and analyze variable relationships through interpretable machine learning to understand key environmental drivers.
Real-time monitoring and rapid identification of drug-resistant pathogenic bacteria risks, save detection time and manpower, understand the environmental drivers of drug-resistant pathogenic bacteria in river systems, help reduce their spread and protect humans and animals. healthy.
Smart Images

Figure CN119068976A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of environmental protection, and in particular to a prediction model construction method for predicting cell density of drug-resistant pathogenic bacteria based on water quality physical and chemical indicators and an application thereof. Background Art
[0002] Pathogenic antibiotic resistant bacteria (PARB) in the environment carry both antibiotic resistance genes (ARGs) and virulence factor genes (VFGs), and thus have both infectious ability and antibiotic resistance, which poses a serious threat to human health and has attracted widespread attention.
[0003] At present, the risk of pathogenic microorganisms caused by drug-resistant pathogens in river water can be reflected by detecting the cell density of drug-resistant pathogens. However, this indicator is not a conventional water quality monitoring indicator and has no standard limit. Its detection method and data processing are relatively complex, especially the need to obtain functional gene data through metagenomic sequencing, identify drug-resistant pathogens through bioinformatics analysis methods, and determine the cell density of water samples through flow cytometry, and then calculate the indicator results through corresponding methods, which cannot meet the real-time monitoring and rapid identification of drug-resistant pathogens. In addition, under multiple stresses in river systems, the environmental driving factors of PARB and their influencing mechanisms are not clear, so it is impossible to effectively reduce the risk of drug-resistant pathogens from the perspective of environmental factor control. Summary of the invention
[0004] In view of the above technical problems, the present invention provides a prediction model construction method for predicting the cell density of drug-resistant pathogenic bacteria based on water quality physical and chemical indicators and its application. The prediction model constructed by the prediction model construction method provided by the present invention has a strong fitting effect, and only requires the water quality physical and chemical indicator test results of the river water sample to be tested to accurately predict the cell density of drug-resistant pathogenic bacteria, and can be used to analyze the response relationship between water quality physical and chemical indicators and river drug-resistant pathogenic bacteria.
[0005] In order to achieve the above-mentioned object of the invention, the present invention adopts the following technical solutions:
[0006] The first aspect of the present invention provides a method for constructing a prediction model for predicting the cell density of drug-resistant pathogenic bacteria based on water quality physical and chemical indicators, which specifically comprises the following steps:
[0007] S1. Take several water samples from the river to be tested, obtain the cell density of drug-resistant pathogenic bacteria, and obtain the water quality physical and chemical index data, the water quality physical and chemical indicators include: carbamazepine (CBZ), ibuprofen (BF), temperature (Temp), 1-chloronaphthalene (1-CN), oxidation-reduction potential (ORP), dissolved oxygen (DO), ammonia nitrogen (NH4 + -N), acenaphthene (Ace), 2,6-dimethylnaphthalene (2,6-DMN), benzo(b)fluoranthene (BbF), benzanthrone (BAT), 2-chloroanthracene (2-ClAnt), dissolved organic carbon (DOC), anthracene (An), total dissolved solids (TDS), pH value, total nitrogen (TN), 2-methylanthraquinone (2-MAQ), electrical conductivity (COND), nitrate nitrogen (NO3 - -N), 1,4-dichloronaphthalene (D-CN), phenanthrene (Phe), atenolol (ATE), naphthalene (Nap), fluoranthene (Flua), benzopyrene (BaP), acenaphthylene (Acy), sulfamethoxazole (SMX), chlorophyll-a (Chl-a), trimethoprim (TMP), benzo(k)fluoranthene (BkF), benzo[g,h,i]perylene (BghiP), total nitrogen / total phosphorus (TN / TP), dissolved organic carbon / total nitrogen (DOC / TN), nitrate nitrogen / total nitrogen (NO3 - -N / TN), ammonia nitrogen / total phosphorus (NH4 + -N / TP);
[0008] S2. Taking the water quality physical and chemical indicators as explanatory variables and the cell density of drug-resistant pathogenic bacteria as the target variable, a regression model is established using a gradient boosting machine (GBM) algorithm to obtain the prediction model.
[0009] The prediction model established by the above prediction model construction method has a strong fitting effect. It only needs the test results of the physical and chemical indicators of the water quality of the river water sample to be tested to accurately predict the cell density of drug-resistant pathogens, so as to achieve real-time monitoring and rapid identification of the risk of drug-resistant pathogens. In addition, the detection of physical and chemical indicators of water bodies is a necessary step in water quality monitoring, and the corresponding detection methods are relatively mature, so this method has strong operability. In addition, the construction method uses automatic machine learning to establish a prediction model, and uses interpretable machine learning to further analyze the relationship between variables. Therefore, it can also be used to analyze the response relationship between water quality physical and chemical indicators and river drug-resistant pathogens, which is conducive to understanding the key environmental driving factors of drug-resistant pathogens in river systems, thereby helping to achieve effective reduction of drug-resistant pathogen risks from the perspective of environmental factor control.
[0010] Preferably, the river to be tested is the Haihe River system.
[0011] Further preferably, the river to be tested is the Chaobai River. The above water quality physical and chemical indicators are more applicable to the prediction of the cell density of drug-resistant pathogenic bacteria in the Chaobai River.
[0012] Preferably, the method for obtaining the cell density of drug-resistant pathogenic bacteria in S1 is: obtaining the relative abundance of drug-resistant pathogenic bacteria by performing metagenomic sequencing on the water sample; detecting the total number of cells in the water sample, dividing the total number of cells by the volume of the water sample to obtain the cell density of the water sample; multiplying the cell density of the water sample by the relative abundance of the drug-resistant pathogenic bacteria to obtain the cell density of the drug-resistant pathogenic bacteria.
[0013] Further preferably, the specific operation of obtaining the relative abundance of the drug-resistant pathogenic bacteria includes:
[0014] Collect water samples, extract metagenomic sequencing data of the samples, and perform quality filtering on raw reads to obtain clean reads (high-quality reads);
[0015] Assembling the clean reads obtained from each sample to obtain contigs (overlapping groups);
[0016] According to the nucleic acid composition and abundance change pattern of the contigs, the sequences are clustered and binned to obtain MAGs (metagenomic assembled genomes, microbial genome drafts), all MAGs are evaluated according to integrity and contamination rate, and high-quality representative MAGs with genome integrity concentration greater than 50% and contamination less than 10% are selected, and the high-quality representative MAGs are de-redundanted to obtain a non-redundant MAGs set;
[0017] Predict ORFs (open reading frames) for all contigs of the non-redundant MAGs set, compare all ORFs with the antibiotic resistance gene database and the virulence factor database in turn, identify MAGs containing both ARG sequences and VFG sequences as potential PARBs, and count the types and quantities of ARGs and VFGs carried by the potential PARBs;
[0018] The abundance of MAGs belonging to the potential PARB in the sample was calculated: the percentage of the number of reads in each sample that could be successfully aligned with all MAGs in the non-redundant MAGs set to the total number of reads in the sample was obtained using a mapping-based method, as well as the coverage of each MAG in the non-redundant MAGs set in each sample. The abundance of each PARB in each sample was calculated according to formula (1):
[0019]
[0020] Among them, the percentage of reads mapping in the sample represents the percentage of the number of reads in a sample that can be successfully aligned with the non-redundant MAGs set to the total number of reads in the sample, the single PARB coverage is the coverage of a MAG belonging to the potential PARB in a sample, and the sum of all MAG coverages represents the sum of the coverage of all MAGs in a sample.
[0021] Further preferably, the total number of cells in the water sample is detected by flow cytometry.
[0022] Preferably, in S2, when the gradient boosting algorithm is used to establish the regression model, the prediction variables are logarithmically transformed.
[0023] Preferably, in S2, the regression model is established by using the h2o package in R software.
[0024] Preferably, S2 also includes testing the prediction ability of the prediction model, and the specific method is:
[0025] The data set of the water sample is randomly divided into a training set and a test set, the training set is used to establish the prediction model, and the test set is used to verify the prediction ability of the prediction model.
[0026] Further preferably, 85% of the data set is used as a training set, and 15% of the data set is used as a test set.
[0027] Further preferably, the fitting coefficient R 2 Measure the predictive power of the forecast model:
[0028] When R 2 When ≤0.3, the fitting effect between the predicted value and the observed value is weak, and the prediction ability of the prediction model is poor;
[0029] When 0.3<R 2 When ≤0.4, the fitting effect between the predicted value and the observed value is weak, and the prediction ability of the prediction model is weak;
[0030] When 0.4<R 2 When ≤0.6, the fitting effect between the predicted value and the observed value is moderate, and the prediction ability of the prediction model is moderate;
[0031] When 0.6<R 2 When ≤1.0, the predicted value and the observed value have a strong fitting effect, and the prediction model has a strong prediction ability.
[0032] The second aspect of the present invention provides the application of the prediction model constructed by the above prediction model construction method in predicting the cell density of drug-resistant pathogenic bacteria.
[0033] The third aspect of the present invention provides the application of the prediction model constructed by the above prediction model construction method in analyzing the response relationship between water quality physical and chemical indicators and river resistant pathogenic bacteria.
[0034] Combined with the third aspect, response relationship analysis is performed through interpretable machine learning methods.
[0035] The fourth aspect of the present invention also provides a method for predicting the cell density of drug-resistant pathogenic bacteria based on physical and chemical indicators of water quality: a prediction model is constructed according to the above-mentioned prediction model construction method, and the prediction model is used to predict the cell density of drug-resistant pathogenic bacteria in the river to be tested.
[0036] Preferably, the operation of predicting the cell density of drug-resistant pathogenic bacteria in the river to be tested includes: obtaining the water quality physical and chemical index data of the river to be tested, substituting it into the prediction model, and obtaining the target variable value, which is the prediction result of the cell density of drug-resistant pathogenic bacteria in the water sample of the river to be tested.
[0037] The beneficial effects of the present invention are as follows: It has been verified that the prediction model established by the prediction model construction method provided by the present invention has a strong fitting effect, and only the test results of the physical and chemical indicators of the water quality of the river water sample to be tested are required to accurately predict the cell density of drug-resistant pathogenic bacteria. On the one hand, it can solve the problem of complex indicator detection and the inability to monitor in real time, which can save a lot of time and manpower and avoid complicated detection processes; on the other hand, it can analyze the response relationship between water quality physical and chemical indicators and river drug-resistant pathogens based on interpretable machine learning methods, which is conducive to understanding the key environmental driving factors of drug-resistant pathogens in river systems, and is crucial to how to reduce the spread of drug-resistant pathogens and protect the health of humans and other animals. The present invention is of great significance to the protection of aquatic organisms and the control of health risks to humans and animals caused by drug-resistant pathogens in water environments. BRIEF DESCRIPTION OF THE DRAWINGS
[0038] Figure 1 It is the model training effect and test effect of predicting the cell density of drug-resistant pathogenic bacteria using water quality physical and chemical index data in the test example of the present invention;
[0039] Figure 2 is a partial dependency relationship diagram between the target variable and each explanatory variable in the prediction model of the cell density of drug-resistant pathogenic bacteria in Example 3 of the present invention;
[0040] Figure 3 It is a SHAP Value diagram between the target variable and each explanatory variable in the prediction model of the cell density of drug-resistant pathogenic bacteria in Example 3 of the present invention. DETAILED DESCRIPTION
[0041] In order to make the purpose, technical scheme and advantages of the present invention clearer, the present invention is further described in detail below in conjunction with specific embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work fall within the scope of protection of the present invention.
[0042] Antibiotic-resistant pathogens (PARB) in the environment have the ability to infect and resist antibiotics, which poses a serious threat to human health. Existing PARB detection methods and data processing are relatively complex, making it impossible to conduct real-time monitoring and rapid identification. In addition, the environmental driving factors of PARB and its influencing mechanisms are not yet clear. Therefore, it is even more difficult to effectively reduce the risk of resistant pathogens through environmental factor control when it is difficult to monitor and quickly identify PARB in real time.
[0043] The present invention establishes a prediction model by taking river water samples in advance and performing automatic machine learning, thereby realizing the prediction of the cell density of drug-resistant pathogenic bacteria in the river water samples to be tested. The method for constructing the prediction model specifically uses the physical and chemical indicator data in water quality monitoring (carbamazepine, ibuprofen, temperature, 1-chloronaphthalene, redox potential, dissolved oxygen, ammonia nitrogen, acenaphthene, 2,6-dimethylnaphthalene, benzo(b)fluoranthene, benzanthrone, 2-chloroanthracene, dissolved organic carbon, anthracene, total dissolved solids, pH value, total nitrogen, 2-methylanthraquinone, conductivity, nitrate nitrogen, 1,4-dichloronaphthalene, phenanthrene, atenolol, naphthalene, fluoranthene, benzopyrene, acenaphthylene, sulfamethoxazole, chlorophyll-a, trimethoprim, benzo(k)fluoranthene, benzo[g,h,i]perylene, total nitrogen / total phosphorus, dissolved organic carbon / total nitrogen, nitrate nitrogen / total nitrogen, ammonia nitrogen / total phosphorus) as explanatory variables, and the cell density of drug-resistant pathogenic bacteria as the predictor variable, and uses the Gradient Boosting Machine (GBM) algorithm to establish the prediction model. Currently, there is no method to predict the cell density of drug-resistant pathogens using physical and chemical indicators of water quality.
[0044] The embodiment of the present invention provides a method for constructing a prediction model for predicting the cell density of drug-resistant pathogenic bacteria based on physical and chemical indicators of water quality.
[0045] The embodiments of the present invention also provide an application of the prediction model constructed by the above prediction model construction method in predicting the cell density of drug-resistant pathogenic bacteria, and provide a method for predicting the cell density of drug-resistant pathogenic bacteria based on physical and chemical indicators of water quality.
[0046] The embodiment of the present invention also provides the application of the prediction model constructed by the above prediction model construction method in analyzing the response relationship between water quality physical and chemical indicators and river drug-resistant pathogenic bacteria.
[0047] The following is further described with reference to specific embodiments.
[0048] If no specific experimental steps or conditions are specified in the following examples, the conventional experimental steps or conditions described in the literature in the art can be used. The raw materials or instruments used are all conventional products that can be obtained commercially, including but not limited to the raw materials or instruments used in the examples of this application.
[0049] Example 1
[0050] This embodiment provides a method for constructing a prediction model for predicting the cell density of drug-resistant pathogenic bacteria based on physical and chemical indicators of water quality.
[0051] (1) Water sample collection
[0052] The Chaobai River (39°-40°N, 116°-117°E) basin in the Haihe River system of China was selected as the research object. Since the Chaobai River has obvious land use zoning characteristics, it was selected as a suitable system to study the impact of human activities on its aquatic ecosystem. Surface water samples (water depth 50 cm) were collected from 34 locations along the river in December 2016, March 2017, June 2017 and September 2017. Among them, samples from seven locations were not collected in December 2016 because the river water formed thick ice at these locations; in June and September 2017, samples from six locations were not collected due to factors such as river drying up.
[0053] 1 mL of surface water was placed in a centrifuge tube for on-site microbial biomass collection and microbial cell quantification using flow cytometry. Approximately 900 mL of surface water was filtered using a 0.45 μm filter membrane with a pump and then LifeGuard was added. TM The solution was stored and then DNA sequencing was performed. All samples were transported to the laboratory on ice.
[0054] A total of 123 water samples were collected. After DNA density testing, it was found that the DNA amount of 9 samples was insufficient for metagenomic sequencing. Therefore, a total of 114 DNA samples were sequenced.
[0055] (2) Testing of water quality physical and chemical indicators
[0056] The 36 water quality physical and chemical indicators used as explanatory variables are: carbamazepine (CBZ), ibuprofen (BF), temperature (Temp), 1-chloronaphthalene (1-CN), oxidation-reduction potential (ORP), dissolved oxygen (DO), ammonia nitrogen (NH4 +-N), acenaphthene (Ace), 2,6-dimethylnaphthalene (2,6-DMN), benzo(b)fluoranthene (BbF), benzanthrone (BAT), 2-chloroanthracene (2-ClAnt), dissolved organic carbon (DOC), anthracene (An), total dissolved solids (TDS), pH value, total nitrogen (TN), 2-methylanthraquinone (2-MAQ), electrical conductivity (COND), nitrate nitrogen (NO3 - -N), 1,4-dichloronaphthalene (D-CN), phenanthrene (Phe), atenolol (ATE), naphthalene (Nap), fluoranthene (Flua), benzopyrene (BaP), acenaphthylene (Acy), sulfamethoxazole (SMX), chlorophyll-a (Chl-a), trimethoprim (TMP), benzo(k)fluoranthene (BkF), benzo[g,h,i]perylene (BghiP), total nitrogen / total phosphorus (TN / TP), dissolved organic carbon / total nitrogen (DOC / TN), nitrate nitrogen / total nitrogen (NO3 - -N / TN), ammonia nitrogen / total phosphorus (NH4 + -N / TP). Among them, the calculation method of the four nutrient ratio indicators is as follows: TN / TP is the ratio of TN to TP, NO3 - -N / TN is NO3 - -N to TN ratio, DOC / TN is the ratio of DOC to TN NH4 + -N / TP is NH4 + -Ratio of N to TP.
[0057] The monitoring methods and corresponding instruments and equipment for the above indicators are shown in Table 1.
[0058] Table 1 Chaobai River water quality monitoring methods and instruments
[0059]
[0060]
[0061] Note: All chemical index test results are the average of three test results.
[0062] (3) Detection of relative abundance of drug-resistant pathogens
[0063] Collect water samples, extract metagenomic sequencing data of the samples, and perform quality filtering on the raw sequencing data (rawreads) to obtain high-quality reads (clean reads);
[0064] Assemble the clean reads obtained from each sample to obtain contigs;
[0065] Further, according to the nucleic acid composition and abundance change pattern of contigs, the sequences were clustered and binned to obtain the metagenomic assembled genomes (MAGs). All MAGs were evaluated according to their integrity and contamination rate, and high-quality representative MAGs with a genome integrity concentration greater than 50% and a contamination rate less than 10% were selected. The high-quality representative MAGs were de-redundanted to obtain a non-redundant MAGs set.
[0066] Predicting open reading frames (ORFs) for all contigs of the non-redundant MAGs set, comparing all ORFs with the antibiotic resistance gene database and the virulence factor database in turn, identifying MAGs containing both ARG sequences and VFG sequences as potential PARBs, and counting the types and quantities of ARGs and VFGs carried by these potential PARBs;
[0067] Calculation of the abundance of potential PARB MAGs in the samples: A mapping-based method was used to obtain the percentage of reads in each sample that could be successfully aligned with all MAGs in the non-redundant MAGs set to the total number of reads in the sample, as well as the coverage of each MAG in the non-redundant gene set in each sample. The abundance of each PARB in each sample was calculated according to formula (1):
[0068]
[0069] Among them, the percentage of reads mapping in the sample represents the percentage of the number of reads in a sample that can be successfully aligned with the non-redundant MAGs set to the total number of reads in the sample, the single PARB coverage is the coverage of a MAG belonging to potential PARB in a sample, and the sum of all MAG coverages represents the sum of the coverage of all MAGs in a sample.
[0070] (4) Detection of cell density of drug-resistant pathogens
[0071] With anhydrous dimethyl sulfoxide Green I was diluted 100 times to prepare the dye, and 10 μL of dye was added to every 100 μL of water sample, and the dye was stained for 15 minutes at room temperature and in the dark. All samples were diluted 10 times with sterile TE buffer (10mMTris-HCl, pH8.0; 1mM EDTA, pH8.0). An equal amount of fluorescent microspheres (BD Company) were added to each sample for quantitative collection of fluorescent signals. The samples were tested for biomass using a flow cytometer to determine the total number of cells in the water sample, which was then divided by the volume of the water sample to obtain the cell density of the water sample (the average of the results of 4 parallel experiments was taken for each sample). The cell density of the resistant pathogens was obtained by multiplying the cell density of the water sample by the relative abundance of the resistant pathogens.
[0072] (5) Using GBM algorithm to establish a prediction model
[0073] The above test data removed some missing data, and the final data set had a total of 114 sets of observations. The 36 water quality physical and chemical indicators (CBZ, BF, Temp, 1-CN, ORP, DO, NH4 + -N, Ace, 2,6-DMN, BbF, BAT, 2-ClAnt, DOC, An, TDS, pH, TN, 2-MAQ, COND, NO3 - -N, D-CN, Phe, ATE, Nap, Flua, BaP, Acy, SMX, Chl-a, TMP, BkF, BghiP, TN / TP, DOC / TN, NO3 - -N / TN, NH4 + -N / TP) as the explanatory variable, and the cell density of drug-resistant pathogens as the target variable, the GBM algorithm was used to establish a regression model through the h2o package in R software. The target variable was logarithmically transformed with a base of 10. The target variable was the cell density of drug-resistant pathogens. When its value was 0, 1 / 10 of the minimum value was taken as the replacement. The parameters ntree and max_depth were determined when the calculation results reached stability. For each model, 85% of the data set (98 samples) was randomly divided into training sets, and the remaining 15% (16 samples) was used as a test set. In addition, the mean square error (MSE) based on the training set and the test set prediction and the fitting coefficient R 2 The two performance parameters are used to evaluate the model. Since different indicators have different numerical ranges, MSE, as an absolute numerical indicator, is only applicable to the comparison of prediction results between the same indicators. In addition, the smaller the MSE, the better the R 2 The larger the value, the larger the R 2 As the main evaluation parameter. 2 When ≤0.3, the predicted value and the observed value have poor fitting effect; when 0.3<R 2≤0.4, the predicted value and the observed value have a weak fitting effect; when 0.4<R 2 When ≤0.6, the predicted value and the observed value have a moderate fitting effect; when 0.6<R 2 When ≤1.0, the predicted value and the observed value have a strong fitting effect. The data set was randomly divided into a training set and a test set using the ratios parameter in the h2o package of R software (v.4.4.0). The model was constructed based on the training set using the h2o.gbm function to obtain a prediction model for the cell density of drug-resistant pathogens, and the generalization ability of the model was tested based on the test set.
[0074] Test example
[0075] The prediction model of Example 1 was tested.
[0076] (1) Model training effect based on training set
[0077] The results of predicting the cell density of drug-resistant pathogens based on the training set using a combination of 36 water quality physical and chemical indicators (explanatory variables) are shown in Table 2. The fitting results of the predicted values and observed values are shown in Figure 1 Gray part.
[0078] Table 2 Training effect of the cell density prediction model of drug-resistant pathogens based on the training set
[0079]
[0080]
[0081]
[0082] As shown in Table 2 and Figure 1 As shown in the figure, the predicted value of the prediction model for the cell density of drug-resistant pathogens in the training set is highly consistent with the detected value, with an MSE of 0.007 and R 2 The value is 0.986, which achieves a strong fitting effect.
[0083] (2) Model prediction effect based on the test set
[0084] The prediction model based on the training set is used in the test set to verify the generalization ability of the model, that is, the prediction ability for new data sets, using the mean square error MSE and the determination coefficient R 2 Conduct an assessment.
[0085] The results of predicting the test set using the drug-resistant pathogen cell density prediction model are shown in Table 3, and the corresponding fitting diagram of the predicted value and the observed value is shown in Figure 1 Black part.
[0086] Table 3 Prediction effect of the cell density prediction model of drug-resistant pathogens based on the test set
[0087] Sample No. Drug-resistant pathogen cell density detection value (after logarithmization) Model prediction value (logarithmized) 1 5.050073 5.218100 2 6.258159 6.340485 3 6.128471 6.176748 4 6.409224 6.178545 5 5.850494 5.685384 6 5.240004 5.144252 7 5.826419 5.847825 8 6.438391 6.646408 9 5.922508 5.940825 10 5.085828 5.011233 11 5.213411 5.239332 12 5.011274 5.457567 13 5.670885 5.397405 14 5.578041 5.966496 15 5.639231 5.490137 16 5.288183 5.64923
[0088] As shown in Table 3 and Figure 1 As shown in Figure 2, the prediction model can still achieve a strong fitting effect (R 2 =0.786), and the MSE is 0.047.
[0089] Example 2
[0090] In this example, Chaobai River is used as the river to be tested, and the prediction model established in Example 1 (based on the training set) is used to predict the cell density of drug-resistant pathogenic bacteria in the river water sample to be tested, and the method is as follows:
[0091] Obtain the water quality physical and chemical index data of the river water sample to be tested, and use the water quality physical and chemical indicators (CBZ, BF, Temp, 1-CN, ORP, DO, NH4 + -N, Ace, 2,6-DMN, BbF, BAT, 2-ClAnt, DOC, An, TDS, pH, TN, 2-MAQ, COND, NO3 - -N, D-CN, Phe, ATE, Nap, Flua, BaP, Acy, SMX, Chl-a, TMP, BkF, BghiP, TN / TP, DOC / TN, NO3 - -N / TN, NH4 + -N / TP) as the explanatory variable, the water quality physical and chemical index data are substituted into the prediction model of drug-resistant pathogenic bacteria cell density, and the obtained target variable value is the prediction result of the drug-resistant pathogenic bacteria cell density of the river water sample to be tested.
[0092] Example 3
[0093] This embodiment provides the application of the prediction model established in Example 1 in analyzing the response relationship between water quality physical and chemical indicators and river drug-resistant pathogens.
[0094] The prediction model of drug-resistant pathogen cell density established in Example 1 was used to create a partial dependency relationship diagram and a SHAP diagram between the target variable and each explanatory variable, as shown in Figure 1. Figure 2 and Figure 3 .
[0095] like Figure 2 and Figure 3 As shown, Figure 2 and Figure 3The order of variables in the partial dependence diagram is ranked by importance, so CBZ and BF are determined to be the main influencing factors of the biomass of drug-resistant pathogens and the primary water quality indicators for the control of drug-resistant pathogens. The PDP diagram between CBZ and PARB shows that there is a nonlinear relationship between these drugs and the cell density of PARB, showing a significant promoting effect in the low concentration range and not significant at high concentrations. In addition, sulfamethoxazole SMX, as an antibiotic drug, also shows a certain positive promoting effect on PARB. In addition, a variety of polycyclic aromatic hydrocarbons and their derivatives also show a positive correlation with the cell density of PARB, such as 1-CN, 2,6-DMN, BbF, BAT, Nap, Ace, etc., indicating that they may have a certain promoting effect on PARB. ORP and PARB show a negative relationship, indicating that PARB is more suitable for growth under reducing conditions and is inhibited under oxidizing conditions. PARB and Chl-a, NH4 + The negative relationship between PARB and NH4+-N indicated that eutrophication of water bodies may inhibit PARB.
[0096] It can be seen from the above experimental results that the prediction model established by the present invention can accurately predict the cell density of drug-resistant pathogenic bacteria in rivers and can be used to monitor the risk of drug-resistant pathogenic bacteria.
[0097] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims
1. A method for constructing a prediction model for predicting the cell density of drug-resistant pathogenic bacteria based on water quality physical and chemical indicators, characterized in that: The specific steps include: S1. Take several water samples from the river to be tested, obtain the cell density of drug-resistant pathogenic bacteria, and obtain the water quality physical and chemical index data, wherein the water quality physical and chemical indicators include: carbamazepine, ibuprofen, temperature, 1-chloronaphthalene, redox potential, dissolved oxygen, ammonia nitrogen, acenaphthene, 2,6-dimethylnaphthalene, benzo(b)fluoranthene, benzanthrone, 2-chloroanthracene, dissolved organic carbon, anthracene, total dissolved solids, pH value, total nitrogen, 2-methylanthraquinone, conductivity, nitrate nitrogen, 1,4-dichloronaphthalene, phenanthrene, atenolol, naphthalene, fluoranthene, benzopyrene, acenaphthylene, sulfamethoxazole, chlorophyll-a, trimethoprim, benzo(k)fluoranthene, benzo[g,h,i]perylene, total nitrogen / total phosphorus, dissolved organic carbon / total nitrogen, nitrate nitrogen / total nitrogen, and ammonia nitrogen / total phosphorus; S2. Taking the water quality physical and chemical indicators as explanatory variables and the cell density of the drug-resistant pathogenic bacteria as the target variable, a regression model is established using a gradient boosting algorithm to obtain the prediction model.
2. The prediction model construction method according to claim 1, characterized in that: The river to be tested is the Haihe River system.
3. The prediction model construction method according to claim 2, characterized in that: The river to be tested is the Chaobai River.
4. The prediction model construction method according to claim 1 or 2, characterized in that: The method for obtaining the cell density of drug-resistant pathogenic bacteria in S1 is: obtaining the relative abundance of drug-resistant pathogenic bacteria by performing metagenomic sequencing on the water sample; detecting the total number of cells in the water sample by a flow cytometer, dividing the total number of cells by the volume of the water sample to obtain the cell density of the water sample; multiplying the cell density of the water sample by the relative abundance of the drug-resistant pathogenic bacteria to obtain the cell density of the drug-resistant pathogenic bacteria.
5. The prediction model construction method according to claim 4, characterized in that: The specific operations for obtaining the relative abundance of the drug-resistant pathogenic bacteria include: Collect water samples, extract metagenomic sequencing data of the samples, and perform quality filtering on raw reads to obtain clean reads; Assembling the clean reads obtained from each sample to obtain contigs; According to the nucleic acid composition and abundance change pattern of the contigs, the sequences are clustered and binned to obtain MAGs, all MAGs are evaluated according to integrity and contamination rate, high-quality representative MAGs with genome integrity concentration greater than 50% and contamination less than 10% are selected, and the high-quality representative MAGs are de-redundanted to obtain a non-redundant MAGs set; Predicting ORFs for all contigs of the non-redundant MAGs set, comparing all ORFs with the antibiotic resistance gene database and the virulence factor database in turn, identifying MAGs containing both ARG sequences and VFG sequences as potential PARBs, and counting the types and quantities of ARGs and VFGs carried by the potential PARBs; The abundance of MAGs belonging to the potential PARB in the sample was calculated: the percentage of the number of reads in each sample that could be successfully aligned with all MAGs in the non-redundant MAGs set to the total number of reads in the sample was obtained using a mapping-based method, as well as the coverage of each MAG in the non-redundant MAGs set in each sample. The abundance of each PARB in each sample was calculated according to formula (1): Among them, the percentage of reads mapping in the sample represents the percentage of the number of reads in a sample that can be successfully aligned with the non-redundant MAGs set to the total number of reads in the sample, the single PARB coverage is the coverage of a MAG belonging to the potential PARB in a sample, and the sum of all MAG coverages represents the sum of the coverage of all MAGs in a sample.
6. The prediction model construction method according to claim 1, characterized in that: Logarithmic transformation of the predictor variables when building the regression model using the gradient boosting algorithm in S2; and / or The regression model was established in S2 using the h2o package in R software.
7. Use of the prediction model constructed by the prediction model construction method according to any one of claims 1 to 6 in predicting the cell density of drug-resistant pathogenic bacteria.
8. Application of the prediction model constructed by the prediction model construction method according to any one of claims 1 to 6 in analyzing the response relationship between water quality physical and chemical indicators and river drug-resistant pathogenic bacteria.
9. A method for predicting the cell density of drug-resistant pathogenic bacteria based on water quality physical and chemical indicators, characterized in that: A prediction model is constructed according to the prediction model construction method according to any one of claims 1 to 6, and the prediction model is used to predict the cell density of drug-resistant pathogenic bacteria in the river to be tested.
10. The method according to claim 9, characterized in that The operation of predicting the cell density of drug-resistant pathogenic bacteria in the river to be tested includes: obtaining the water quality physical and chemical index data of the river to be tested, substituting it into the prediction model, and obtaining the target variable value, which is the prediction result of the cell density of drug-resistant pathogenic bacteria in the water sample of the river to be tested.
Citation Information
Patent Citations
Method for predicting physicochemical indexes of river water quality by using bacterial OTUs abundance
CN114841411A
Method for predicting growth condition of microcystis based on conventional water quality physicochemical indexes
CN115472230A
River ecosystem health assessment method based on microbial community specific response
WO2023098239A1
