Machine learning enhanced silver-based covalent organic framework nano-enzyme colorimetric detection method and application thereof

Abstract

The invention discloses a machine learning enhanced silver-based covalent organic framework nano-enzyme colorimetric detection method and application thereof, and constructs a nano-enzyme which is formed by loading silver nanoparticles on a covalent organic framework and has oxide-like enzyme activity, and the nano-enzyme is used for rapid visual detection of mercury ions. When mercury ions exist, silver ions in the COF-AgNPs can form silver amalgam with the mercury ions, so that a colorless substrate 3, 3 ', 5, 5'-tetramethyl benzidine is catalyzed to generate a blue oxidation product oxTMB. And the mercury content can be judged by detecting the absorbance and RGB value change of the chromogenic reaction. And meanwhile, a machine learning data analysis strategy of a support vector machine model is adopted to carry out five-fold layered cross validation analysis on the obtained absorbance value and RGB value, so that the detection range is expanded, and the detection sensitivity is improved.

Description

Technical Field

[0001] The present invention relates to a machine learning enhanced silver-based covalent organic framework nanoenzyme colorimetric detection method, which can detect mercury ions (Hg 2+ ) for rapid detection, which belongs to the research fields of analytical chemistry, food safety and testing, and environmental analysis and monitoring. Background Art

[0002] Mercury ions (Hg 2+ ) As a type of highly toxic heavy metal pollutant, it has become a major threat to the global water ecological environment and food safety due to its high bioaccumulation and environmental persistence. With the intensification of industrial wastewater discharge, mineral mining and agricultural activities, mercury ions enter aquatic products (such as fish, shellfish, shrimp, etc.) through surface runoff and bioaccumulation, eventually causing irreversible damage to the human nervous system, kidneys and immune system. The World Health Organization (WHO) and the Minamata Convention have listed mercury pollution as a priority control target and strictly limited the maximum residual limit of mercury in aquatic products. It is urgent to develop efficient and accurate detection technologies to ensure food safety and public health. Traditional detection of Hg 2+ Current technologies include atomic fluorescence spectrometry (AFS) and inductively coupled plasma mass spectrometry (ICP-MS). While highly sensitive, these rely on large instruments, complex sample pretreatment, and specialized procedures, making them difficult to meet the demands of rapid on-site screening. Furthermore, the complex matrix of aquatic products can easily lead to signal offsets, requiring multiple purification steps to eliminate interference. This is not only time-consuming but can also introduce secondary contamination.

[0003] In recent years, nanozymes have shown great application potential in the fields of food safety, disease diagnosis and treatment, environmental monitoring and biocatalysis due to their convenient synthesis, low cost, excellent stability and easy storage and transportation. In particular, noble metal nanozymes have become the research focus in the field of colorimetric detection due to their high surface functionalization ability, excellent biocompatibility and biomimetic catalytic performance. In particular, silver-based nanozymes have received increasing attention in the field of catalysis due to their lower cost and easier preparation. For example, the Sun team constructed a mercury ion (Hg) nanozyme by embedding small-sized silver nanoparticles (AgNPs) into a temperature-responsive gelatin matrix. 2+ ) colorimetric sensor with specific catalytic activation, which can realize Hg in complex media such as blood and wastewater 2+high-throughput detection of mercury; Yin et al. innovatively used carbon nitride-doped melamine-silver system to endow the material with dual functional properties of peroxidase activity and visible light catalytic ability, and simultaneously achieved total mercury detection and photocatalytic detoxification. However, these nanozymes inevitably have the problem of poor catalytic stability, which hinders their wider application. Therefore, in response to the above-mentioned catalytic stability problem, many scientists have tried to prepare nanozymes with high catalytic stability by combining various organic compounds with precious metal nanoparticles (NPs) for the detection of mercury ions (Hg 2+ ) analysis and detection. In these studies, covalent organic frameworks (COFs) have shown significant advantages due to their unique structural properties. Covalent organic frameworks (COFs), as a new type of crystalline porous polymer formed by strong covalent bonds between elements such as carbon (C), nitrogen (N), and oxygen (O), have become ideal carriers for loading precious metal nanoparticles due to their high specific surface area, adjustable pore size and excellent stability. The strong interaction between its nitrogen-containing groups and nanoparticles can not only improve the stability of the material, but also control the size of the nanoparticles through the pore confinement effect, thereby enhancing the catalytic activity and sensitivity. For example, Wang's team developed a method for Hg based on the Cu2O@Cu2S core-shell D-TA-COF heterostructure. 2+ The biosensor with high sensitivity and dual-modal detection capability has realized the detection of Hg in complex environments (such as river water and serum). 2+ The Li team developed a method for the analysis of trace amounts of Hg based on two-dimensional covalent organic framework nanosheets (COFPTAzo) loaded with gold nanoparticles (AuNPs). 2+ The colorimetric platform, which combines ultrasensitive response and excellent selectivity, fully demonstrates the potential of functional nanocomposites to enhance sensing performance and environmental adaptability through organic-inorganic collaborative design.

[0004] In order to further break through the mercury ion (Hg 2+) detection, the linear response range of a single colorimetric sensor is narrow, and it is difficult to balance the detection of low and high concentrations. Machine learning provides an innovative solution for nanoscale enzyme sensing technology. As a powerful data analysis tool, machine learning is based on statistical, probabilistic and computational theory to build a data-driven algorithm system. It has significant advantages in processing complex data, classification and prediction. Through supervised learning (such as support vector machine, random forest), unsupervised learning (such as principal component analysis), etc., it can independently mine potential laws from multi-dimensional data and build prediction models. In the field of sensing technology and nanomaterials, machine learning technology effectively breaks through the response limitations and selectivity bottlenecks of traditional single sensors by integrating multi-source data such as ultraviolet spectra and RGB images. Taking the Prussian blue nanoscale enzyme multi-modal biosensor as an example, this platform innovatively integrates colorimetric and photothermal dual signal detection mechanisms, combines convolutional neural network (CNN) algorithm and smartphone terminal, not only realizes high-sensitivity portable detection of biomarkers, but also expands the detection dynamic range to continuous coverage from trace (pM level) to high concentration (μM level) through the unique non-linear modeling advantage of machine learning. This 'nanoscale enzyme + machine learning' collaborative strategy not only overcomes the interference problem in complex matrix of traditional methods, but also improves the prediction accuracy across concentration intervals through algorithm optimization, providing an intelligent detection new paradigm with high precision and practicality for environmental monitoring and clinical diagnosis. Therefore, according to the progress of the related research of predecessors, combining the advantages of COF-AgNPs nanoscale enzyme and machine learning, a kind of colorimetric sensor based on nanoscale enzyme for Hg 2+ detection is constructed. SUMMARY

[0005] In view of the technical problems existing in the detection of mercury ions (Hg 2+ ) in aquatic products in the background art, a machine learning enhanced nanoscale enzyme colorimetric sensor for rapid detection of mercury ions (Hg 2+ ) in aquatic products is proposed. Compared with traditional mercury ion (Hg 2+ ) detection methods such as atomic fluorescence spectrometry, cold atomic absorption spectrometry and inductively coupled plasma mass spectrometry, the method uses covalent organic framework (COF) loaded silver nanoparticles (AgNPs) as nanoscale enzyme (COF-AgNPs) probe to detect mercury ions (Hg 2+ ) in different aquatic products. Based on the strong selectivity of AgNPs to mercury ions (Hg 2+ ), the nanoscale enzyme colorimetric technology solves the problem of mercury ions (Hg 2+) selectivity problem, and the formation of silver-mercury alloy makes the sensitivity and detection limit have certain advantages over traditional methods; then the classification performance of five machine learning models including random forest (RF), support vector machine (SVM), K nearest neighbor (KNN), logistic regression (LR) and linear discriminant analysis (LDA) were systematically compared, as well as the classification capabilities under spectral single mode, RGB single mode and fusion multi-mode. It not only greatly improves the detection of mercury ions (Hg) by nanozyme colorimetry technology in different aquatic products, but also greatly improves the detection of mercury ions (Hg) by nanozyme colorimetry technology in different aquatic products. 2+ ) detection, and also realizes the detection of mercury ions (Hg 2+ ) Rapid test score.

[0006] In order to achieve the above object, the technical solution adopted by the present invention is as follows: A machine learning-enhanced silver-based covalent organic framework nanozyme colorimetric detection method, the method is as follows: S1: The COF-AgNPs composite material, acetate buffer, and 3,3',5,5'-tetramethylbenzidine (TMB) were mixed at a volume ratio of 0.02:10:2 to obtain a sample solution; S2: Add 100 μL of mercury ions Hg with a concentration range of 0.05-50 μmol / L to the sample solution obtained in S1. 2+ Standard solution to obtain test solution; S3: The mercury ions Hg obtained in S2 with different concentrations 2+ The reaction time of the standard solution test solution in a 30°C environment is 5 minutes; S4: Measure the absorbance of the test solution obtained in S3 within the range of 400-800 nm, and add mercury ions Hg 2+ The difference in absorbance at 652 nm between the test solution before and after the standard solution is related to the mercury ion Hg 2+ A linear model was constructed based on the concentration of mercury ions. A smartphone was used to obtain the RGB three-channel color values ​​of the LED light source simultaneously. The grayscale value was calculated using the formula Gray = 0.299R + 0.587G + 0.114B, and the grayscale value was compared with the mercury ion Hg 2+ Concentration to build a linear regression model; S5: The collected samples contain different concentrations of mercury ions Hg 2+ The absorbance values ​​of the test solution in the range of 400-800 nm were pre-treated with multivariate scattering correction (MSC) to eliminate light scattering interference; S6: After the data preprocessed in S5, competitive adaptive reweighted sampling (CARS) is used for feature extraction; S7: The feature values ​​extracted in S6 and the RGB three-channel color values ​​extracted from the test liquid are used to construct spectral single-mode, RGB single-mode and fusion dual-mode datasets; S8: Five classification and recognition models were constructed based on the collected spectral single-mode, RGB single-mode, and fusion dual-mode data sets, and different classification and recognition models were evaluated to obtain the best model.

[0007] Furthermore, the synthesis method of the COF-AgNPs composite material is as follows: COFs are dispersed in anhydrous ethanol at a ratio of 1:1 and ultrasonically mixed for 1 h, polyvinylpyrrolidone PVP with Mw≈5800 is added at a mass ratio of 3:1, and 2.5 mL of 0.1 mol / L silver nitrate, sodium citrate and glucose solution are added in sequence under ultrasonic assistance, and stirred for 15 h to successfully load AgNPs on the COFs surface. The yellow solid is collected by centrifugation and washed several times with a mixture of pure water and methanol, and finally vacuum dried to obtain the COF-AgNPs composite material.

[0008] Furthermore, the synthesis method of the COFs is as follows: 1,3,5-tris(4-aminophenyl)benzene and 2,5-divinylterephthalaldehyde DVA are placed in a conical flask at a molar mass ratio of 2:3, 50 mL of acetonitrile containing 2.9 mol / L acetic acid HAc is added, and after ultrasonication, the mixture is placed at room temperature for 72 h, centrifuged and washed several times with a mixture of tetrahydrofuran (THF) and anhydrous ethanol (EtOH), and finally vacuum dried to obtain yellow powdered COFs.

[0009] Furthermore, the five types of classification and recognition models are: Random Forest (RF), Support Vector Machine (SVM), K-Nearest Neighbors (KNN), Logistic Regression (LR) and Linear Discriminant Analysis (LDA).

[0010] Furthermore, the indicators of the initial screening evaluation of the model include accuracy, recall, precision and F1 Fraction;

[0011]

[0012]

[0013]

[0014] Among them, Accuracy is the accuracy rate, Recall is the recall rate, Precision is the precision rate and F1is a score, TP is a true positive example, FN is a false negative example, FP is a false positive example, and TN is a true negative example.

[0015] Furthermore, the rules of competitive adaptive reweighted sampling CARS are: Monte Carlo sampling times 100 times, cross-validation coefficient 5, and feature retention coefficient 11.

[0016] Furthermore, the support vector machine (SVM) model was the best model with an accuracy of 95.97%. F1 The value was 0.958, the detection range was extended to a continuous linear interval of 0–50 μmol / L, and the detection limit of the nanozyme colorimetric sensor was 0.0107 μmol / L.

[0017] A machine learning-enhanced silver-based covalent organic framework nanozyme colorimetric detection method for mercury ions (Hg 2+ ) for rapid testing.

[0018] The above technical solution can achieve the following beneficial effects: The innovative detection method proposed in this paper based on the combination of covalent organic framework-loaded silver nanoparticles (COF-AgNPs) composite nanozymes and multimodal machine learning has successfully achieved the detection of mercury ions (Hg 2+ ) in a wide linear range of 0-50 μmol / L with full coverage detection and concentration partitioning optimization. This method uses 2,5-divinylterephthalaldehyde and 1,3,5-tris(4-aminophenyl)benzene as monomers to synthesize hierarchical flower-like covalent organic frameworks (COFs) in an acetonitrile / glacial acetic acid catalytic system, and constructs COF-AgNPs nanozyme probes by in situ growth of silver nanoparticles (AgNPs). When mercury ions (Hg 2+ ) or mercury-containing compounds, Hg 2+ Through metal replacement reaction (Ag 0 + Hg 2+ → Ag + + Hg 0 ) in situ forms an Ag-Hg alloy on the surface of AgNPs. The synergistic effect of its electronic structure significantly reduces the activation energy of TMB oxidation reaction. At the same time, the bimetallic synergistic sites jointly promote the generation of reactive oxygen species (singlet oxygen ( 1 O2), superoxide anion (O2 •−)), achieving efficient oxidation of TMB to generate blue oxTMB. A dual-mode detection system of absorbance signal and RGB color value was successfully established. In order to meet the needs of accurate detection in a wide concentration range, the classification performance of five machine learning models, including random forest (RF), support vector machine (SVM), K nearest neighbor (KNN), logistic regression (LR) and linear discriminant analysis (LDA), was systematically evaluated under spectral single modality, RGB single modality and multimodal fusion conditions. The experimental results show that the multimodal data fusion strategy significantly improves the model's ability to analyze complex concentration gradients by integrating the quantitative characteristics of colorimetry with the response advantages of RGB numerical analysis. This strategy not only breaks through the detection range limitations of traditional single-modality methods, but also significantly improves data processing efficiency and result accuracy through collaborative analysis of multi-source signals, providing an innovative technical solution for high-throughput monitoring of heavy metal pollutants in the food and environmental fields.

[0019] Nanozyme probes based on in situ growth of AgNPs on hierarchical flower-like COFs, whose porous structure provides a high specific surface area and nitrogen-rich group structure, enhance the interaction between AgNPs and mercury ions (Hg 2+ ) binding capacity. Experimental verification showed that the spiked recovery of this method in actual fish and shrimp samples was 90.18%-103.219%, and the relative standard deviation (RSD) was only 0.9%-8.87%, indicating that it has significant resistance to interference from complex matrices. At the same time, RGB image data acquisition is compatible with low-cost portable devices (smartphone cameras), breaking through the dependence of traditional ICP-MS, AAS and other instruments on laboratory environments, and providing feasibility for rapid on-site screening.

[0020] Through the dual signal acquisition mechanism of colorimetry (ultraviolet absorption spectrum) and RGB numerical analysis, combined with KNN, SVM, RF and other machine learning algorithms to build a multimodal fusion model, the detection of mercury ions (Hg 2+ ). Its average accuracy is 91.98%, and the detection limit is as low as 0.01μmol / L, which is more than 10% more sensitive than traditional spectrophotometry. This breakthrough is due to the synergistic effect of spectral characteristics and color channel values, which effectively overcomes the limitations of a single modality for complex samples. In addition, the differentiated model optimization strategy for different concentration ranges (low, medium, and high) has made the detection accuracy of the lowest concentration range (0.01-1μmol / L) as high as 95.08%, solving the industry difficulty that low-concentration signals in trace heavy metal detection are easily interfered by noise. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 Nanomaterial synthesis route (1A), mercury ions (Hg 2+ ) detection flowchart (1B) and machine learning processing flowchart (1C); Figure 2For different mercury ions (Hg 2+ ) concentration and the characteristic peak intensity at 500-800 nm in the UV absorption spectrum (2A) and linear equation diagrams (2B, 2C) and different mercury ions (Hg 2+ ) linear equation relationship diagram between concentration and RGB grayscale value (2D); Figure 3 Performance comparison of different models built for multimodality and unimodality (3A) and classification performance comparison of each model (3B); Figure 4 Confusion matrix (4A) and receiver operating characteristic curve (4B) generated for the support vector machine condition; Figure 5 SHAP feature contribution graph; Figure 6 To interfere with the experimental results. DETAILED DESCRIPTION

[0022] The following is combined with Figure 1-5 The present invention will be further described with examples: A machine learning-enhanced colorimetric detection method for silver-based covalent organic framework nanozymes was developed. The COFs were synthesized as follows: 1,3,5-tris(4-aminophenyl)benzene and 2,5-divinylterephthalaldehyde (DVA) were placed in a conical flask at a molar ratio of 2:3. 50 mL of acetonitrile containing 2.9 mol / L acetic acid (HAc) was added. After sonication, the mixture was incubated at room temperature for 72 hours. The mixture was centrifuged and washed three times with a mixture of tetrahydrofuran (THF) and anhydrous ethanol (EtOH). Finally, the COFs were vacuum dried to obtain a yellow powder.

[0023] The COF-AgNPs composite was synthesized as follows: the COFs obtained above were dispersed in anhydrous ethanol at a ratio of 1:1 and ultrasonically mixed for 1 hour. Polyvinylpyrrolidone (PVP) (Mw ≈ 5800) was then added at a mass ratio of 3:1. 2.5 mL of 0.1 mol / L silver nitrate, sodium citrate, and glucose solutions were then added sequentially under ultrasound assistance and stirred for 15 hours to successfully load the AgNPs onto the COF surface. The yellow solid was collected by centrifugation and washed three times with a mixture of pure water and methanol. Finally, the COF-AgNPs composite was obtained by vacuum drying.

[0024] A machine learning model construction method for a machine learning-enhanced silver-based covalent organic framework nanozyme colorimetric detection method is as follows: S1: The COF-AgNPs composite material, acetate buffer, and 3,3',5,5'-tetramethylbenzidine (TMB) were mixed at a volume ratio of 0.02:10:2 to obtain a sample solution; S2: Add 100 μL of mercury ions Hg with a concentration range of 0.05-50 μmol / L to the sample solution obtained in S1. 2+ Standard solution to obtain test solution; S3: The mercury ions Hg obtained in S2 with different concentrations 2+ The reaction time of the standard solution test solution in a 30°C environment is 5 minutes; S4: Measure the absorbance of the test solution obtained in S3 within the range of 400-800 nm, and add mercury ions Hg 2+ The difference in absorbance at 652 nm between the test solution before and after the standard solution is related to the mercury ion Hg 2+ A linear model was constructed based on the concentration of mercury ions. A smartphone was used to obtain the RGB three-channel color values ​​of the LED light source simultaneously. The grayscale value was calculated using the formula Gray = 0.299R + 0.587G + 0.114B, and the grayscale value was compared with the mercury ion Hg 2+ Concentration to build a linear regression model; S5: The collected samples contain different concentrations of mercury ions Hg 2+ The absorbance values ​​of the test solution in the range of 400-800 nm were pre-treated with multivariate scattering correction (MSC) to eliminate light scattering interference; S6: After the data preprocessed in S5, competitive adaptive reweighted sampling (CARS) is used for feature extraction; S7: The feature values ​​extracted in S6 and the RGB three-channel color values ​​extracted from the test liquid are used to construct spectral single-mode, RGB single-mode and fusion dual-mode datasets; S8: Five classification and recognition models were constructed based on the collected spectral single-mode, RGB single-mode, and fusion dual-mode datasets. The model construction types included random forest (RF), support vector machine (SVM), K-nearest neighbor (KNN), logistic regression (LR), and linear discriminant analysis (LDA). Different classification and recognition models were evaluated to obtain the optimal model.

[0025] Example 1: Mercury ions (Hg 2+ ) Detection and analysis results Mercury ions (Hg 2+ ) Standard solution preparation: Prepare mercury ion (Hg 2+ Weigh 0.162 g of mercuric nitrate and dissolve it in 10 mL of nitric acid. Transfer the solution to a 1000 mL volumetric flask and dilute to the mark with deionized water.

[0026] Detection process: Different concentrations of mercury ions (Hg 2+) standard solution, acetate buffer, and chromogenic substrate 3,3',5,5'-tetramethylbenzidine were mixed in a volume ratio of 1:10:2, and then the COF-AgNPs solution was added to obtain a test sample solution; after reacting at 30°C for 5 minutes, the test sample solution was subjected to absorbance detection in the range of 400-800 nm and the RGB values ​​were synchronized with a smartphone (iPhone Xs) + LED light source (standard D65 (5500 k)).

[0027] Constructing linear models: Constructing different Hg 2+ The relationship curve between concentration and characteristic peak intensity at 500-800 nm in the UV absorption spectrum is as follows: Figure 2 As shown in A; by linear fitting, in the range of 0.05-0.5 μmol / L and 5-50 μmol / L, ΔA652 and c(Hg 2+ ) showed a good linear relationship (R 2 >0.99), such as Figure 2 As shown in B, the linear regression equation in the range of 0.05-0.5 μmol / L is Y=0. 4398 c(Hg 2+ )-0.03468, (R 2 =0.990). According to the 3σ criterion, the limit of detection (LOD) was calculated to be 0.0107 μmol / L. Figure 2 The linear regression equation in the range of 5-50 μmol / L shown in C is Y=0.0217c(Hg 2+ )+0.2858,(R 2 =0.9937), and the detection limit was 0.207 μmol / L. RGB data acquisition was performed under a standard D65 (5500 k) light source, and grayscale values ​​were extracted through image processing to construct a correlation between different mercury ions (Hg 2+ ) concentration, such as Figure 2 D, linear fitting uses the gray value formula: Gray=0.2999R+0.587G+0.114B, the detection range is 0-50 µmol / L, through linear fitting, in the range of 0.5-15 µmol, the gray value is consistent with c(Hg 2+ ) showed a good linear relationship (R 2 >0.99). The equation is Y=-3.29323 c(Hg 2+ )+153.1088,(R 2=0.995). Based on the 3σ criterion, the calculated limit of detection (LOD) was approximately 1.32 μmol / L. In summary, the detection limit of the nanozyme sensor was 0.0107 μmol / L (approximately 0.002 mg / L), which is much lower than the Chinese national standard (GB 2762-2022) and the EU mercury residue limit for aquatic products listed in Table 1. Furthermore, the sensor exhibited a wider linear range and lower detection limit than other sensors (Table 2).

[0028] Table 1 Detection standards for mercury in aquatic products

[0029] Table 2 Comparison of other methods for detecting mercury ions

[0030] The collected UV absorption spectrum data were corrected for baseline drift and scattering interference by multivariate scattering correction (MSC), and the competitive adaptive reweighted sampling (CARS) algorithm was used to optimize the spectral features (Monte Carlo iteration 100 times, 5-fold cross validation, retaining 11 key wavelengths). At the same time, the mean values ​​of the R, G, and B channels of the color reaction solution were extracted to construct a visual feature set. Then, the spectral features (10 dimensions), color features (3 dimensions) and their fusion features (13 dimensions) were input into the random forest (RF), support vector machine (SVM), K nearest neighbor (KNN), logistic regression (LR) and linear discriminant analysis (LDA) models for training. The accuracy, recall, precision and F1 Multi-dimensional evaluation of scores. A comprehensive comparison of different models ( Figure 3 A) and different modes ( Figure 3 B) The classification performance of the model shows that the spectral-color fusion dataset combined with the support vector machine (SVM) model shows the best performance with an accuracy of 95.97%. F1 The value is 0.958, which is more than 10% higher than the single spectrum model. We selected the support vector machine (SVM) model to generate the confusion matrix, receiver operating characteristic (ROC) curve and area under the curve (AUC) to further analyze the sensor performance. Figure 4 As shown in A, the confusion matrix shows that the biosensor performs better at low and high concentrations, with an overall accuracy of 93.58%, but there is still room for improvement at medium concentrations. Figure 4 In B), the AUC values ​​corresponding to low concentration (0.99), medium concentration (0.96) and high concentration (0.97) are all at a high level, which shows that the biosensor can effectively distinguish Hg-containing 2+ Hg-free 2+Among them, the biosensor performs best at high concentration (black curve), followed by low concentration (yellow curve), and slightly weaker at medium concentration (green curve), but the overall performance is still good.

[0031] SHAP interpretability analysis ( Figure 5 The results (shown in Figure 3) further validate that the synergistic effect of absorbance and RGB channel values ​​dominates the prediction results (contribution ratios of 62% and 38%, respectively), providing a quantitative basis for the multimodal sensing data fusion mechanism. Furthermore, Table 3 shows that by calibrating the signal using the algorithm, the detection range is expanded to a continuous linear interval of 0–50 µmol / L, covering the entire concentration range of traditional methods and eliminating blind spots in the intermediate intervals, significantly improving the dynamic range and applicability of the assay.

[0032] Table 3 Comparison of methods

[0033] High selectivity for mercury ions (Hg 2+ ) is crucial and is an important indicator of whether the method can be used in actual sample testing. Since seafood still contains some metal ions after digestion, it may affect the detection of mercury ions (Hg 2+ ) selectivity, so the metal ions and anions commonly found in seafood (Al 3+ , Pb 2+ Mg 2 + 、Cd 2+ , Ca 2+ 、Mn 2+ 、SO4 2- 、CO3 2- 、Cl - PO4 3- , K + 、Zn 2+ 、Na + 、Ni 4+ 、Cu 2+ ), the metal ion concentration is mercury ion (Hg 2 + ) is 10 times that of mercury ions (Hg 2+ ) was tested for selectivity at 10 times the level of mercury ions (Hg 2+ ) Under the optimal conditions for detection, different metal ions were added to the same system and the absorbance at 652 nm was measured using an enzyme marker. Figure 6 It can be seen that the removal of mercury ions (Hg 2+ ), other ions responded little to the colorimetric detection method, indicating that the method is not very sensitive to mercury ions (Hg 2+) has high specificity and will not be interfered by other ions.

[0034] In order to demonstrate the feasibility of this method in analytical applications, five kinds of seafood, including shrimp, squid, yellow catfish, clams and kelp, were selected as representative samples to evaluate the analytical performance of the proposed method. The seafood samples needed to be digested first. The specific steps are as follows: Weigh 5 kinds of seafood (0.5 g) and place them in conical flasks respectively. Add 15 mL HNO3 and 2.5 mL H2SO4 in sequence. The resulting mixed solution was digested at room temperature overnight. After that, the mixture in the 5 conical flasks was heated and boiled until a large amount of white smoke was produced and the solution became clear. After cooling to room temperature, the pH of the obtained clear solution was adjusted to 5.0 with 1 mol / L NaOH solution and then diluted to 100 mL with deionized water. Under the optimal detection conditions, different concentrations of Hg were added to the solutions of different seafood samples. 2+ The standard solution was spiked and recovered, and mercury ions (Hg 2+ ) standard solution was added in a volume of 100 μL. The absorbance at 652 nm was measured using a microplate reader, and the RGB values ​​were captured using a smartphone. Spike recovery was calculated based on the established standard curve. All experiments were repeated three times.

[0035] Table 4 Actual sample test table

[0036] All experimental conditions were consistent with the aqueous solution conditions. As shown in Table 4, the spike recovery rate was between 90.18% and 103.219%, and the relative standard deviation (RSD) was between 0.9% and 8.87%. The above results show that mercury ions (Hg 2+ )The colorimetric detection method for enhancing the oxidase activity of COF-AgNPs is feasible and reliable, with satisfactory accuracy and precision, and has broad application prospects.

[0037] The above are all preferred embodiments of the present invention. For ordinary technicians in this technical field, without departing from the principle of the present invention, various equivalent modifications to the present invention are within the scope of protection of the claims attached to this application.

Claims

1. A machine learning-enhanced silver-based covalent organic framework nanozyme colorimetric detection method, characterized in that: The synthesis method of the covalent organic framework is as follows: first, 1,3,5-tris(4-aminophenyl)benzene and 2,5-divinylterephthalaldehyde (DVA) were added to an acetonitrile solution containing 2.9 mol / L acetic acid (HAc) in a 2:3 molar ratio. After ultrasonic dispersion, the mixture was allowed to react for 72 hours. The yellow product was obtained by centrifugation, washed several times with tetrahydrofuran and anhydrous ethanol, and then dried in vacuum. The obtained COFs were then dispersed with anhydrous ethanol in a 1:1 ratio, and polyvinylpyrrolidone (PVP) equivalent to 3 times the mass of the COFs was added as a stabilizer. 2.5 mL of 0.1 mol / L silver nitrate, sodium citrate, and glucose solutions were injected in sequence. Silver nanoparticles (AgNPs) were loaded on the COFs surface by ultrasonic-assisted stirring for 15 hours. Finally, the yellow complex was collected by centrifugation, washed with pure water and methanol, and dried in vacuum to obtain the COF-AgNPs composite material. The method is as follows: S1: mixing the COF-AgNPs composite material according to claim 1, acetate buffer, and 3,3',5,5'-tetramethylbenzidine (TMB) in a volume ratio of 0.02:10:2 to obtain a sample solution; S2: Add 100 μL of mercury ions Hg with a concentration range of 0.05-50 μmol / L to the sample solution obtained in S1. 2+ Standard solution to obtain test solution; S3: The mercury ions Hg obtained in S2 with different concentrations 2+ The reaction time of the standard solution test solution in a 30°C environment is 5 minutes; S4: Measure the absorbance of the test solution obtained in S3 within the range of 400-800 nm, and add mercury ions Hg 2+ The difference in absorbance at 652 nm between the test solution before and after the standard solution is related to the mercury ion Hg 2+ A linear model was constructed based on the concentration of mercury ions. A smartphone was used to obtain the RGB three-channel color values ​​of the LED light source simultaneously. The grayscale value was calculated using the formula Gray = 0.299R + 0.587G + 0.114B, and the grayscale value was compared with the mercury ion Hg 2+ Concentration to build a linear regression model; S5: The collected samples contain different concentrations of mercury ions Hg 2+ The absorbance values ​​of the test solution in the range of 400-800 nm were pre-treated with multivariate scattering correction (MSC) to eliminate light scattering interference; S6: After the data preprocessed in S5, competitive adaptive reweighted sampling (CARS) is used for feature extraction; S7: The feature values ​​extracted in S6 and the RGB three-channel color values ​​extracted from the test liquid are used to construct spectral single-mode, RGB single-mode and fusion dual-mode datasets; S8: Five classification and recognition models were constructed based on the collected spectral single-mode, RGB single-mode, and fusion dual-mode data sets, and different classification and recognition models were evaluated to obtain the best model.

2. The machine learning-enhanced silver-based covalent organic framework nanozyme colorimetric detection method according to claim 1, characterized in that: The five classification and recognition models are: random forest RF, support vector machine SVM, K nearest neighbor KNN, logistic regression LR and linear discriminant analysis LDA.

3. The machine learning-enhanced silver-based covalent organic framework nanozyme colorimetric detection method according to claim 1, characterized in that: The indicators of the initial screening evaluation of the model include Accuracy, Recall, Precision and F1 Fraction; ； ； ； ； Among them, Accuracy is the accuracy rate, Recall is the recall rate, Precision is the precision rate and F1 is a score, TP is a true positive example, FN is a false negative example, FP is a false positive example, and TN is a true negative example.

4. The machine learning-enhanced silver-based covalent organic framework nanozyme colorimetric detection method according to claim 1, characterized in that: The rules of competitive adaptive reweighted sampling CARS are: 100 Monte Carlo sampling times, 5 cross-validation coefficient, and 11 feature retention coefficient.

5. A machine learning enhanced silver-based covalent organic framework nanozyme colorimetric detection method according to claim 1 or 3, characterized in that: The support vector machine (SVM) model is the best model with an accuracy of 95.97%. F1 The value was 0.958, the detection range was extended to a continuous linear interval of 0–50 μmol / L, and the detection limit of the nanozyme colorimetric sensor was 0.0107 μmol / L.

6. A machine learning-enhanced silver-based covalent organic framework nanozyme colorimetric detection method for mercury ions Hg in food and environmental samples 2+ An application for rapid testing.

Images