Rapid Method for Comprehensive Toxicity Measurement of Sewage
A predictive model using zebrafish embryo data addresses inefficiencies in conventional sewage toxicity measurement by integrating traditional and behavioral indices, enabling rapid and accurate comprehensive toxicity assessment with reduced embryo use and high model accuracy.
Patent Information
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2024-11-08
- Publication Date
- 2026-04-02
AI Technical Summary
Conventional methods for measuring sewage toxicity using zebrafish embryos are inefficient and limited to neurotoxicity indicators, requiring multiple experiments and filtered samples, which do not reflect actual sewage conditions.
A rapid method using zebrafish embryo toxicity indicators and behavioral data to establish a predictive model for comprehensive sewage toxicity, incorporating traditional and behavioral toxicity indices, and selecting appropriate models based on LC50 and LC10 values to determine overall toxicity.
Enables rapid, accurate, and cost-effective measurement of sewage toxicity with reduced embryo usage, maintaining high model accuracy (R² > 0.85) and providing comprehensive toxicity indicators.
Smart Images
Figure 2026057423000001_ABST
Abstract
Description
[Technical Field]
[0001] This invention relates to the field of water quality monitoring, and more particularly to a rapid method for comprehensively measuring the toxicity of sewage. [Background technology]
[0002] Human activities release many pollutants into the environment, potentially posing risks to human health and the environment. While traditional chemical methods are commonly used to measure specific pollutants in sewage, sewage often exists as a complex mixture, making chemical analysis unable to identify many components and increasing uncertainty in water quality assessment. Compared to chemical analysis, biological analysis can comprehensively measure the actual toxicity of all pollutants, i.e., the overall toxicity of sewage. Zebrafish embryos are widely used in biological analysis due to their rapid reproductive cycle and transparent embryos. Comprehensive toxicity indicators obtained by conventional experimental methods are LC (Low-Coefficient of Tolerance). 10 or LC 50 (Units are percentages of water sample concentration) and this method is standardized and widely recognized. However, this method requires setting multiple concentration gradients and conducting multiple parallel experiments, which increases the amount of embryos used and the experimental operation time. To solve this problem, Chinese patent application No. 202010577907.X discloses a method for evaluating the developmental neurotoxicity of wastewater. This method has the advantage of being simple, fast, and efficient, as it involves exposing transgenic zebrafish embryos to concentrated or diluted wastewater up to 24 hpf, removing the exposed embryos, observing them with a fluorescence inverted microscope, and collecting images for analysis. However, because the water samples measured by this method need to be filtered and concentrated, the measurement results may not reflect the actual state of the sewage, and the results obtained by this method are only neurotoxicity indicators, not comprehensive sewage toxicity. [Overview of the project] [Problems that the invention aims to solve]
[0003] Objective of the Invention: The present invention aims to provide a method for rapidly measuring the comprehensive toxicity of sewage. This method not only allows for obtaining standardized comprehensive toxicity indicators for actual sewage samples, but also maintains the advantages of speed, ease of operation, and low cost, making it suitable for rapid measurement of the comprehensive toxicity of large quantities of actual sewage samples. [Means for solving the problem]
[0004] Technical solution: The rapid measurement method for comprehensive toxicity of sewage according to the present invention includes the following steps.
[0005] Step (S1) Perform data preprocessing by measuring zebrafish embryo toxicity indicators of sewage samples in the full process of several sewage treatment plants, including traditional toxicity indicators and behavioral toxicity indicators.
[0006] Step (S2): Establish a comprehensive wastewater toxicity prediction model based on different target variables, with traditional toxicity indicators as target variables and behavioral toxicity indicators as feature variables.
[0007] Step (S3) involves inputting zebrafish embryo behavioral toxicity index data of the sewage sample for measurement into a predictive model, selecting a corresponding predictive model according to the prediction result of the target variable, and obtaining the comprehensive toxicity of the sewage sample for measurement.
[0008] Preferably, the traditional toxicity index is the median lethal dose (LC). 50 and 10% lethal dose LC 10 This is obtained by including and observing the mortality rate of zebrafish embryos within a specific time period.
[0009] Preferably, in step (S2), the comprehensive wastewater toxicity prediction model based on the different target variables includes the following:
[0010] Depending on the target variable, LC 50 Select <100 sewage samples and calculate the median lethal dose (LC). 50A prediction model 1 is obtained by training with as the target variable, where LC 50 Select sewage samples with LC 50 ≥ 100 and obtain a prediction model 2 by training with the 10% lethal dose LC 10 as the target variable.
[0011] Preferably, in step (S3), obtaining the comprehensive toxicity of the sewage sample for measurement by selecting the corresponding prediction model according to the prediction result of the target variable includes the following.
[0012] Input the embryo behavioral toxicity index data of the sewage sample for measurement as input data into the prediction model. When the output value of prediction model 1 is LC 50 < 100, the median lethal dose LC 50 is the comprehensive toxicity of the sewage sample for measurement.
[0013] When the output value of prediction model 1 is LC 50 ≥ 100, input the embryo behavioral toxicity index data of the sewage sample for measurement as input data into prediction model 2, and obtain the output value LC 10 of prediction model 2, that is, obtain the comprehensive toxicity of the sewage sample for measurement.
[0014] Preferably, the behavioral toxicity index includes the activity of zebrafish embryos in a dark environment within a specific time, the activity of zebrafish embryos in a bright environment within a specific time, the total distance of movement of zebrafish embryos in a dark environment within a specific time, the total distance of movement of zebrafish embryos in a bright environment within a specific time, the explosion distance of zebrafish embryos within a specific time, the cruising distance of zebrafish embryos within a specific time, and the freezing distance of zebrafish embryos within a specific time, and is measured by a zebrafish behavior detector.
[0015] Preferably, a Lasso model is used to establish the sewage comprehensive toxicity prediction model.
[0016] Preferably, in the establishment of the sewage comprehensive toxicity prediction model, the coefficient of determination R 2The performance of the comprehensive sewage toxicity prediction model is evaluated based on the coefficient of determination R 2 As a method of calculation, TIFF2026057423000002.tif1751 However, TIFF2026057423000003.tif55 is the coefficient of determination. TIFF2026057423000004.tif55 is the value of the i-th traditional toxicity index in the test set. TIFF2026057423000005.tif65 is the mean value of traditional toxicity indicators within the test set. TIFF2026057423000006.tif65 is the predicted value output by the prediction model, and n is the number of traditional toxicity index values.
[0017] Preferably, the data preprocessing includes data cleaning, data standardization, and feature selection.
[0018] Preferably, data standardization utilizes the Z-Score method or the Max-Min method.
[0019] Preferably, the feature selection method utilizes the Pearson correlation coefficient measurement method. [Effects of the Invention]
[0020] Beneficial Effects: Compared to conventional techniques, the present invention has the following notable advantages: 1. The present invention enables rapid and easy toxicity measurement of large quantities of wastewater samples, significantly reducing the number of embryos used compared to conventional experiments; 2. The present invention further improves the accuracy of the model in actual use by selecting the corresponding predictive model according to the prediction results of the target variable; 3. The model coefficient of determination R established in the present invention. 2 Because it is >0.85, the accuracy is high. [Brief explanation of the drawing]
[0021] [Figure 1] This is a flowchart of the method of the present invention. [Modes for carrying out the invention]
[0022] The embodiments of the present invention will be further described below with reference to the attached drawings.
[0023] As shown in Figure 1, the rapid measurement method for comprehensive toxicity of sewage according to the present invention mainly comprises a step (1) of establishing a comprehensive toxicity model for sewage and a step (2) of measuring comprehensive toxicity of sewage.
[0024] Of these, the stage of establishing a comprehensive toxicity model for sewage (1) mainly includes the following steps.
[0025] (1.1) Data collection: Measurement of zebrafish embryo toxicity indices using sewage samples from the full process at several sewage treatment plants. Zebrafish embryo toxicity indices include traditional toxicity indices and behavioral toxicity indices.
[0026] Of these, the traditional toxicity indicator is the median lethal dose (LC). 50 and 10% lethal dose LC 10 This is obtained by observing the mortality rate of zebrafish embryos within a specific time period.
[0027] Behavioral toxicity indicators, measured by a zebrafish behavior detector, include the activity of zebrafish embryos in a dark environment for a specified period of time (activity-dark), the activity of zebrafish embryos in a bright environment for a specified period of time (activity-light), the total distance traveled by zebrafish embryos in a dark environment for a specified period of time (total distance-dark), the total distance traveled by zebrafish embryos in a bright environment for a specified period of time (total distance-light), the explosion distance of zebrafish embryos in dark and bright environments for a specified period of time, the cruising distance of zebrafish embryos in dark and bright environments for a specified period of time, and the freezing distance of zebrafish embryos in dark and bright environments for a specified period of time.
[0028] (1.2) Data preprocessing includes data cleaning, data standardization, and feature selection.
[0029] The aforementioned data standardization includes, but is not limited to, the Z-Score method and the Max-Min method. In this embodiment, the Z-Score method was used, but as a calculation method, TIFF2026057423000007.tif1951
[0030] however, TIFF2026057423000008.tif55 is the standard score, TIFF2026057423000009.tif66 is a value for one of the behavioral toxicity indicators, TIFF2026057423000010.tif65 is the average value of the relevant indicator data. TIFF2026057423000011.tif55 is the standard deviation of the index data.
[0031] The aforementioned feature selection method utilized the Pearson correlation coefficient measurement method, but the correlation coefficient r of the two indicators 2 If >0.8, remove one of the indicators and calculate the correlation coefficient between each pair of features. 2 Set it to <0.8.
[0032] The processed data is divided into a training set and a test set.
[0033] (1.3) Establishing the Model We will establish a comprehensive wastewater toxicity prediction model based on the Lasso model, using traditional toxicity indicators as target variables and behavioral toxicity indicators as features.
[0034] (1.4) Model training and testing The predictive model is trained using the training set data, and then the coefficient of determination R is calculated using the test set. 2 The performance of the predictive model is evaluated based on this.
[0035] Of these, the coefficient of determination R 2 As a method of calculation, TIFF2026057423000012.tif1651
[0036] however, TIFF2026057423000013.tif55 is the coefficient of determination. TIFF2026057423000014.tif55 is the value of the i-th traditional toxicity index in the test set. TIFF2026057423000015.tif65 is the mean value of traditional toxicity indicators within the test set. TIFF2026057423000016.tif65 is the i-th predicted value output based on the features in the test set, where n is the number of traditional toxicity index values.
[0037] Specifically, depending on the target variable, LC 50 Select a sewage sample with a concentration <100 (i.e., a water sample concentration of 100%) and perform LC (Low-Cooling) testing. 50 Predictive model 1 was obtained by training with LC as the target variable, 50 Select a sewage sample with a value of ≥100 and perform LC. 10 Predictive model 2 was obtained by training with the target variable.
[0038] The second stage of establishing a comprehensive toxicity model for sewage mainly includes the following:
[0039] (2.1) Measurement of behavioral toxicity index of zebrafish embryos using water samples for measurement: The behavioral toxicity index of the wastewater samples for measurement is measured using zebrafish embryos, and the data is preprocessed.
[0040] (2.2) The pre-processed behavioral toxicity index data is input to Prediction Model 1 as input data, and the output value LC of Prediction Model 1 is obtained. 50 If <100, the LC 50 The value represents the overall toxicity of the water sample.
[0041] The output value of prediction model 1 is LC 50 If the value is ≥100, the embryonic behavioral toxicity index data of the sewage sample used for measurement is input to prediction model 2, and the output value LC of prediction model 2 is obtained. 10That is, the cerebrotoxicity of the sewage sample was obtained.
[0042] The method of the present invention will be further described below using embodiments.
[0043] The sewage samples used to establish the comprehensive sewage toxicity prediction model were collected from multiple process sections of 40 sewage treatment plants, totaling 300 samples. The sewage samples used for measurement were actual influent and actual wastewater from a certain urban sewage treatment plant. The rapid measurement method for comprehensive sewage toxicity according to the present invention includes the following steps.
[0044] (1) LC based on behavioral data of zebrafish embryos 50 and LC 10 Establish a predictive model for this.
[0045] (1.1) Data collection: Zebrafish embryo toxicity indicators were measured from 300 sewage samples, including traditional toxicity indicators and behavioral toxicity indicators. Traditional toxicity indicators included LC, which is measured by observing embryo mortality within a specific time period. 50 and LC 10 This includes, however, in this embodiment, LC 50 <100 sewage samples, 130, LC 50 There are 170 sewage samples with a tolerance of ≥100. Behavioral toxicity indicators, including embryo activity in a 5-minute dark environment, embryo activity in a 5-minute bright environment, total distance traveled in a 5-minute dark environment, total distance traveled in a 5-minute bright environment, explosion distance in both dark and bright environments for 5 minutes, cruising distance, and freezing distance, were measured using a zebrafish behavior detector.
[0046] (1.2) Data preprocessing: Data cleaning, data standardization, and feature selection are performed on the collected data. The Z-score method is used for data standardization, and the Pearson correlation coefficient method is used for feature selection to obtain the final correlation coefficient r between each pair of features. 2 The range is 0.496-0.770.
[0047] (1.3) Model Training: The dataset is divided into a 70% training set and a 30% test set. The Lasso model is trained with traditional toxicity indicators as target variables and behavioral toxicity indicators as features. LC 50 <100 water samples are 130, LC 50 Since the water sample with a temperature of ≥100 is 170, LC is used to obtain prediction model 1. 50 Using the target variable, 130 water samples were collected using LC to obtain predictive model 2. 10 170 water samples were trained with this as the target variable.
[0048] (1.4) Model testing: Based on the trained model, use the data from the test set to determine the coefficient of determination R 2 The performance of the model is evaluated based on this. Of these, Predictive Model 1: R 2 =0.874, and prediction model 2:R 2 This equals 0.893.
[0049] (2) Conduct comprehensive toxicity measurements of sewage using established models.
[0050] (a) The sewage sample for measurement is the actual inflow water from the urban sewage treatment plant. (2.1a) Measurement of behavioral toxicity index: The behavioral toxicity index of the sewage samples was measured using zebrafish embryos. Ten parallel groups were set up for each water sample, and the obtained behavioral toxicity index data are shown in Table 1.
[0051] [Table 1]
[0052] (2.2a) Model selection: Predictive model 1 is input with the mean values of the obtained behavioral toxicity indicators as input data, and influent water LC 50 =65 was obtained. LC 50 Since <100, LC 50 =65 represents the overall toxicity of the water sample.
[0053] (b) The wastewater sample used for measurement is actual wastewater from the municipal wastewater treatment plant. (2.1b) Measurement of behavioral toxicity index: The behavioral toxicity index of the sewage samples was measured using zebrafish embryos. Ten parallel groups were set up for each water sample, and the obtained behavioral toxicity index data are shown in Table 2.
[0054] [Table 2]
[0055] (2.2b) Model selection: The average value of the obtained behavioral toxicity index is input to predictive model 1 as input data, and the LC of the effluent is calculated. 50 The result obtained was =136. LC of the effluent 50 >100 indicates that the toxicity of the water sample used for measurement is low, LC 50 This means that it is not suitable to use it as a toxicity indicator, so the behavioral toxicity indicator parameters of the runoff water are input into prediction model 2, and the comprehensive toxicity value of the water sample is LC. 10 I obtained = 43.
Claims
1. Step (S1) Perform data preprocessing by measuring zebrafish embryo toxicity indicators of sewage samples in the full process of a sewage treatment plant, including traditional toxicity indicators and behavioral toxicity indicators. Step (S2), establish a comprehensive wastewater toxicity prediction model based on different target variables, with the traditional toxicity index as the target variable and the behavioral toxicity index as the characteristic feature. Step (S3): Input the data of the zebrafish embryo behavioral toxicity index of the sewage sample for measurement as input data into the prediction model, select the corresponding prediction model according to the prediction result of the target variable, and obtain the comprehensive toxicity of the sewage sample for measurement. A rapid method for comprehensively measuring the toxicity of sewage, characterized by including the following step.
2. The aforementioned traditional toxicity indicators are based on the median lethal dose (LC). 50 and 10% lethal dose LC 10 A rapid method for measuring the comprehensive toxicity of sewage as described in 1, characterized by including and obtaining the result by observing the mortality rate of zebrafish embryos within a specific time period.
3. The comprehensive wastewater toxicity prediction model based on the different target variables in step (S2) includes LC by the different target variables. 50 <Select 100 sewage samples and determine the median lethal dose LC.> 50 Predictive model 1 obtained by training with the target variable, and LC 50 Select a sewage sample with a value of ≥100 and use the 10% lethal dose LC. 10 A rapid method for measuring the comprehensive toxicity of sewage as described in 2, characterized by including a predictive model 2 obtained by training with a target variable.
4. In step (S3) above, selecting a corresponding prediction model according to the prediction result of the target variable to obtain the comprehensive toxicity of the wastewater sample for measurement is: The embryonic behavioral toxicity index data of the measured sewage sample is input to the prediction model 1 as input data, and the output value of the prediction model 1 is LC 50 If <100, the median lethal dose (LC) 50 This becomes the holistic toxicity of sewage. When the output value of the prediction model 1 is LC 50 ≥ 100, the embryotoxicity index data of the sewage sample for measurement is input into the prediction model 2 as input data, and the 10% lethal dose LC 10 which is the output value of the obtained prediction model 2, becomes the comprehensive toxicity of the sewage sample for measurement A rapid method for measuring the comprehensive toxicity of sewage as described in claim 3, characterized by the above.
5. The rapid method for measuring the comprehensive toxicity of sewage as described in paragraph 1, characterized in that the behavioral toxicity index includes the activity of zebrafish embryos in a dark environment for a specified period of time, the activity of zebrafish embryos in a bright environment for a specified period of time, the total distance traveled by zebrafish embryos in a dark environment for a specified period of time, the total distance traveled by zebrafish embryos in a bright environment for a specified period of time, the explosion distance of zebrafish embryos for a specified period of time, the cruising distance of zebrafish embryos for a specified period of time, and the freezing distance of zebrafish embryos for a specified period of time, and is measured by a zebrafish behavior detector.
6. The rapid measurement method for the comprehensive toxicity of sewage described in 1, characterized in that the Lasso model is used to establish the aforementioned comprehensive sewage toxicity prediction model.
7. In establishing the aforementioned comprehensive sewage toxicity prediction model, the coefficient of determination R 2 The performance of the comprehensive sewage toxicity prediction model is evaluated based on the coefficient of determination R 2 As a method of calculation, however, This is the coefficient of determination, This is the value of the i-th traditional toxicity index in the test set. This represents the average value of traditional toxicity indicators within the test set. The rapid measurement method for comprehensive toxicity of sewage described in 1, characterized in that n is a predicted value output from a prediction model, and n is the number of traditional toxicity index values.
8. The rapid measurement method for the comprehensive toxicity of sewage as described in 1, characterized in that the data preprocessing includes data cleaning, data standardization, and feature selection.
9. The rapid measurement method for comprehensive toxicity of sewage as described in 8, characterized in that the data standardization utilizes the Z-Score method or the Max-Min method.
10. The method for selecting the aforementioned features is characterized by utilizing the Pearson correlation coefficient measurement method, and is a rapid method for measuring the comprehensive toxicity of sewage as described in 8.