Optimization method for detecting a2 protein content in milk based on elisa technology

Abstract

The application discloses an A2 protein content detection optimization method in milk based on an ELISA technology, relates to the technical field of protein content detection, and comprises the following steps: collecting milk matrix interference detection data sets, establishing a milk matrix interference corrector, generating an ELISA dynamic detection module, preparing a milk sample to be detected, controlling the ELISA dynamic detection module to perform protein optimization detection on the milk sample to be detected, obtaining milk absorbance detection data, establishing an absorbance-concentration standard curve, and determining a milk A2 protein content detection result. The application solves the technical problem that, in the prior art, when the ELISA detects the A2 protein in milk, the detection is prone to component interference, and the adaptability to different types of milk is insufficient, resulting in poor detection accuracy and reliability. The application achieves more accurate detection of the A2 protein content in milk and stronger anti-interference capability, and improves the technical effect of the accuracy and reliability of the ELISA technology in detecting the A2 protein in milk.

Description

Technical Field

[0001] This invention relates to the field of protein content detection technology, specifically to an optimized method for detecting A2 protein content in milk based on ELISA technology. Background Technology

[0002] In the field of A2 protein content detection in milk, ELISA technology is widely used due to its advantages such as high specificity and relatively simple operation. However, its detection results are easily affected by the complex components of the milk matrix. Matrix factors such as fat, casein polymers, and Maillard reaction products in milk can lead to deviations in detection results by affecting the specific binding of antigen and antibody and interfering with absorbance signal reading. Furthermore, existing ELISA detection methods mostly use fixed detection parameters and procedures, making it difficult to adapt to the characteristics of different types of milk samples such as whole milk, skim milk, pasteurized milk, and UHT sterilized milk. For example, the high fat content of whole milk and the low fat content of skim milk have different requirements for incubation efficiency and washing effect in the detection steps. Fixed parameters can easily lead to insufficient detection accuracy for some samples, thus limiting the application effect of ELISA technology in the accurate detection of A2 protein content in milk.

[0003] Existing technologies have technical problems when using ELISA to detect A2 protein in milk. These problems include susceptibility to component interference and insufficient adaptability to different types of milk, resulting in poor accuracy and reliability. Summary of the Invention

[0004] This application provides an optimized method for detecting A2 protein content in milk based on ELISA technology, which addresses the technical problems of poor accuracy and reliability in the detection of A2 protein in milk by existing ELISA technology, which is easily affected by component interference and has insufficient adaptability to different types of milk.

[0005] In view of the above problems, this application provides an optimized method for detecting A2 protein content in milk based on ELISA technology.

[0006] This application provides an optimized method for detecting A2 protein content in milk based on ELISA technology. The method includes:

[0007] A milk matrix interference detection dataset was collected, and a correction training fit was performed based on the dataset to establish a milk matrix interference corrector. An ELISA adaptive detection module was constructed, and the milk matrix interference corrector was coupled to the ELISA adaptive detection module to generate an ELISA dynamic detection module. Milk samples to be tested were prepared, and the ELISA dynamic detection module was controlled to perform protein optimization detection on the milk samples to obtain milk absorbance detection data. An absorbance-concentration standard curve was established, and the protein content was converted based on the absorbance-concentration standard curve to determine the milk A2 protein content detection result.

[0008] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0009] A milk matrix interference detection dataset was collected, and a milk matrix interference corrector was established. An ELISA adaptive detection module was constructed, and the milk matrix interference corrector was coupled to this module to generate a dynamic ELISA detection module. Milk samples were prepared for testing, and protein optimization detection was performed on these samples to obtain milk absorbance data. An absorbance-concentration standard curve was established, and the protein content was converted based on this curve to determine the milk A2 protein content. This method achieves more accurate and interference-resistant detection of A2 protein content in milk, improving the accuracy and reliability of ELISA technology for detecting A2 protein in milk. Attached Figure Description

[0010] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0011] Figure 1 A schematic flowchart illustrating the optimized method for detecting A2 protein content in milk based on ELISA technology provided in this application embodiment;

[0012] Figure 2 This is a schematic diagram of the process for collecting the milk matrix interference detection dataset in the optimized method for detecting A2 protein content in milk based on ELISA technology provided in the embodiments of this application. Detailed Implementation

[0013] This application provides an optimized method for detecting A2 protein content in milk based on ELISA technology, which addresses the technical problems of existing ELISA methods for detecting A2 protein in milk, such as susceptibility to component interference and insufficient adaptability to different types of milk, resulting in poor detection accuracy and reliability.

[0014] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0015] Examples, such as Figure 1 As shown, this application provides an optimized method for detecting A2 protein content in milk based on ELISA technology, the method comprising:

[0016] Step S100: Collect a milk matrix interference detection dataset, perform correction training and fitting based on the milk matrix interference detection dataset, and establish a milk matrix interference corrector.

[0017] Specifically, establishing a milk matrix interference corrector requires a two-stage process: The first stage involves collecting a milk matrix interference detection dataset. First, based on milk A2 protein detection technology, the set of milk matrix interference factors is identified. Horizontal gradient analysis is performed on each interference factor within the set to determine the horizontal gradient set of matrix interference factors. Then, based on this gradient set, orthogonal experiments are designed to construct a matrix interference factor test parameter table. Finally, ELISA technology is used to conduct milk matrix interference tests according to the parameter table, completing the dataset collection. The second stage involves correcting and training the corrector based on the dataset. First, the matrix interference factor dataset and the corresponding milk absorbance influence dataset are extracted from the collected dataset. The former is used as the input variable, and the latter as the output variable, and labels are used to form a milk matrix interference sample set. Then, a support vector machine is used to train and fit the absorbance of the sample set to obtain an initial matrix interference corrector. Finally, cross-validation and iterative optimization of the initial corrector are used to ultimately establish a milk matrix interference corrector that can effectively counteract milk matrix interference.

[0018] Step S200: Construct an ELISA adaptive detection module, and couple the milk matrix interference corrector to the ELISA adaptive detection module to generate an ELISA dynamic detection module.

[0019] Specifically, the first step is to construct an ELISA adaptive detection module. This involves obtaining the complete ELISA detection process and extracting detection nodes to form an ELISA detection node set. Next, each detection node in the node set is broken down and analyzed, transforming the operational logic and parameter requirements of each node into an executable ELISA node detection program. Then, for each node detection program, the impact of different control parameters on the detection results is analyzed, and a set of key control parameters for each node program is selected. Combined with a set of milk sample types, the key control parameter sets are analyzed according to sample type to determine the threshold values ​​for node control strategy parameters corresponding to different milk types. Finally, ELISA milk protein detection targets are preset, such as detection accuracy and efficiency requirements. Guided by these targets, the control strategy parameters are optimized within the threshold range of node control strategy parameters for each milk type. The optimized milk type-target node control strategy parameter identifier is stored in the detection module, completing the construction of the ELISA adaptive detection module. Based on this, the established milk matrix interference corrector is functionally coupled with the constructed ELISA adaptive detection module, so that when the adaptive detection module performs the detection operation, it can synchronously call the milk matrix interference corrector to correct the absorbance deviation caused by matrix interference in real time, and finally generate an ELISA dynamic detection module with both adaptive adjustment and interference correction capabilities.

[0020] Step S300: Prepare a milk sample to be tested, and control the ELISA dynamic detection module to perform protein optimization detection on the milk sample to be tested, and obtain milk absorbance detection data.

[0021] Specifically, the process begins with the preparation of the milk sample to be tested. A protein detection pretreatment procedure is constructed, incorporating low-temperature defatting, protein denaturation inhibition, and multi-stage filtration. The target milk sample is then obtained and subjected to the pretreatment procedure, including low-temperature defatting to remove fat interference, protein denaturation inhibition to preserve A2 protein activity, and multi-stage filtration to remove impurities, resulting in a usable milk sample. This usable milk sample is then homogenized to ensure uniform composition before being aliquoted and stored to prevent spoilage or compositional changes, ultimately preparing a milk sample that meets the testing requirements. Next, the ELISA dynamic detection module is controlled for optimized protein detection. First, the target milk type corresponding to the milk sample is identified and obtained. Then, based on this target milk type, the ELISA dynamic detection module is controlled to match the corresponding node control strategy parameters, such as the reaction temperature and incubation time, to suit the milk type. Through the module's built-in adaptive detection function and milk matrix interference corrector, protein optimization detection is performed on the sample, correcting the impact of matrix interference on the detection results in real time. Finally, milk absorbance detection data reflecting the A2 protein content in the sample is output.

[0022] Step S400: Establish an absorbance-concentration standard curve, and convert the absorbance detection data of milk into protein content based on the absorbance-concentration standard curve to determine the detection result of milk A2 protein content.

[0023] Specifically, an absorbance-concentration standard curve was established. First, standard A2 protein milk samples with different concentration gradients were prepared. Then, ELISA technology was used to detect these standard samples and read the data, simultaneously acquiring the corresponding standard milk A2 protein concentration data and standard milk absorbance data. Subsequently, using the standard milk A2 protein concentration data as the x-axis and the standard milk absorbance data as the y-axis, a correlation model between the two was constructed through logistic regression fitting, ultimately establishing an absorbance-concentration standard curve reflecting the relationship between absorbance and A2 protein concentration. Based on this, protein content conversion and result determination were performed: First, based on the established absorbance-concentration standard curve, the obtained milk absorbance detection data was substituted into the curve model to complete the conversion from absorbance to A2 protein concentration, obtaining initial milk A2 protein content detection data. Then, a spike recovery test was performed on the initial detection data. By adding standard A2 protein to samples with known concentrations, the accuracy of the detection method was verified. Based on the test results, deviation corrections were made to the initial data, ultimately determining accurate and reliable milk A2 protein content detection results.

[0024] In one possible implementation, such as Figure 2 As shown, step S100 further includes:

[0025] Step S110: Obtain the set of milk matrix interference factors based on milk A2 protein detection technology.

[0026] Step S120: Perform horizontal gradient analysis on each interference factor in the milk matrix interference factor set to determine the matrix interference factor horizontal gradient set.

[0027] Step S130: Based on the horizontal gradient set of the matrix interference factors, perform orthogonal experimental design and construct a test parameter table for the matrix interference factors.

[0028] Step S140: Use ELISA technology to perform milk matrix interference test based on the matrix interference factor test parameter table, and collect the milk matrix interference detection dataset.

[0029] Specifically, the initial step in collecting the milk matrix interference detection dataset revolves around the principles and practical detection scenarios of milk A2 protein detection technology. The implementation is as follows: First, it is clarified that the milk A2 protein detection technology is based on ELISA, whose detection process relies on the binding reaction between the antigen milk A2 protein and a specific antibody. The absorbance value reflects the binding amount, which in turn correlates with the A2 protein content. Based on this technical logic, milk matrix components and properties that may interfere with this binding reaction or absorbance reading are identified, ultimately forming a milk matrix interference factor set. This set covers key interfering factors in milk that may affect the accuracy of detection, such as milk fat content, lactose concentration, non-target proteins like A1 protein and casein, which may bind non-specifically to antibodies, and the pH value of the milk itself. This lays the foundation for subsequent gradient analysis and testing of interfering factors.

[0030] For the acquired set of interfering factors in the milk matrix, including fat content, lactose concentration, non-target protein content, and pH value, the natural distribution range and possible fluctuation range of each interfering factor in actual milk samples were analyzed one by one. For example, the common range of milk fat content is approximately 0.5% to 4.0%, pH value is usually between 6.5 and 6.8, and lactose concentration is generally 4.5% to 5.0%. Then, based on the sensitivity of ELISA detection to interfering factors, several representative gradient levels were defined for each interfering factor to ensure that the gradient can cover... Its main range of variation can accurately capture the differences in the impact of different levels on the test results. For example, the fat content can be set with gradients of 0.5%, 1.0%, 2.0%, 3.0%, and 4.0%, and the pH value can be set with gradients of 6.4, 6.5, 6.6, 6.7, 6.8, and 6.9. Finally, all interfering factors and their corresponding gradient levels are integrated to form a matrix interfering factor level gradient set containing information such as the name of each interfering factor, gradient range, and specific gradient value. This provides a clear basis for variables and levels for the subsequent construction of test parameter tables based on orthogonal experiments.

[0031] The core objective of orthogonal experimental design is to reduce redundant tests and improve data acquisition efficiency while ensuring coverage of all interfering factors and their interactions. Then, based on a defined set of matrix interfering factor level gradients, all interfering factors to be investigated, such as fat content, lactose concentration, non-target protein content, and pH value, along with the number of gradient levels for each factor, are identified. A suitable orthogonal array is selected; for example, if four interfering factors are involved and each factor has five gradient levels, an L25 orthogonal array can be used. Next, according to the row and column structure of the orthogonal array, the different gradient levels of each interfering factor are sequentially assigned to the corresponding columns and rows, forming an experimental scheme that includes combinations of interfering factors and gradient levels, ensuring that all level combinations of every two interfering factors are evenly distributed. Finally, based on this experimental scheme, basic operational parameters required for ELISA detection are added, such as incubation temperature, incubation time, and number of washes. The interfering factor gradient combinations and basic operational parameters are integrated into a structured table, constructing a complete matrix interfering factor test parameter table.

[0032] Based on the constructed matrix interference factor test parameter table, corresponding simulated milk matrix samples were prepared one by one according to the gradient combination of interference factors and the basic operating parameters set for each group of interference factors in the table. That is, for each group of parameters, milk matrix with specific gradients of fat content, lactose concentration, non-target protein content, and pH value was prepared, and A2 protein standard of known concentration was added to ensure that the A2 protein content of the samples was consistent except for matrix interference factors, so as to accurately reflect the impact of matrix interference on the detection results. Next, the standard detection procedure of ELISA technology was strictly followed, and the detection operation was performed on the corresponding simulated milk matrix samples according to the basic operating parameters in each group of parameters, such as incubation temperature, incubation time, and number of washes: sample loading, primary antibody incubation, washing to remove unbound antibodies, secondary antibody incubation, washing again, colorimetric reaction, and finally reading the absorbance value of each group of samples using a microplate reader. During this process, the complete parameter information corresponding to each test was recorded simultaneously, including the specific gradient values ​​of each interference factor, basic operating parameters, and final absorbance detection data, to ensure that the parameters and data corresponded one-to-one. Finally, the absorbance data of all groups of interference factors were integrated and archived to form a complete milk matrix interference detection dataset.

[0033] In one possible implementation, step S100 further includes:

[0034] Step S150: Based on the milk matrix interference detection dataset, obtain the matrix interference factor dataset and the corresponding milk absorbance influence dataset.

[0035] Step S160: Using the matrix interference factor dataset as the input variable and the corresponding milk absorbance influence dataset as the output variable, the input variable and the output variable are divided and labeled to obtain the milk matrix interference sample set.

[0036] Step S170: Use a support vector machine to train and fit the absorbance correction of the milk matrix interference sample set to obtain an initial matrix interference corrector.

[0037] Step S180: Perform cross-validation and iterative optimization on the initial matrix interference corrector to establish the milk matrix interference corrector.

[0038] Specifically, the composition of the milk matrix interference detection dataset is defined. This dataset contains complete information recorded from previous tests conducted using ELISA technology according to the matrix interference factor test parameter table. It covers the specific parameter values ​​of each matrix interference factor in each test group, such as fat content, lactose concentration, non-target protein content, and pH value, as well as the absorbance data of milk samples read by a microplate reader under the corresponding test conditions. Based on this, the dataset is categorized and split: On the one hand, the matrix interference-related parameter information is extracted from all test groups, integrating the gradient values ​​of fat content, lactose concentration, non-target protein content, and pH value for each test group to form a matrix-based matrix interference factor dataset. This dataset directly reflects the interference conditions under different test scenarios. On the other hand, the absorbance detection results corresponding to each test group are extracted. Since these absorbance data were obtained under specific matrix interference conditions, they can reflect the influence of interference factors on the detection signal. Therefore, they are organized according to their correspondence with the interference factor parameters to form a milk absorbance influence dataset. Ultimately, this ensures that each combination of interference parameters in the matrix interference factor dataset accurately corresponds to a specific absorbance value in the milk absorbance influence dataset, laying the data foundation for subsequently constructing an input-output correlation sample set.

[0039] The variable correspondences were clearly defined, and the obtained matrix interference factor dataset was used as the input variable for the subsequent model correction. This dataset contains interference parameters such as fat content, lactose concentration, non-target protein content, and pH value for each test group, which are the core influencing factors causing absorbance changes. The corresponding milk absorbance impact dataset was used as the model's output variable. This dataset records the absorbance detection results for each set of interference parameters, directly reflecting the degree of influence of interference factors on the detection signal. Next, a one-to-one correspondence identification process was performed on the input and output variables: a unique identifier was assigned to each set of input variables, i.e., a specific combination of interference parameters, and the corresponding output variable was associated with the same identifier, i.e., the absorbance value under the same test scenario, ensuring that the correspondence between each set of interference parameter combinations and absorbance results is clear and traceable. Finally, all input and output variables with unified identifiers were integrated according to their identifiers to form a structured milk matrix interference sample set. This sample set contains both the feature data required for model training, i.e., input variables, and the training target data, i.e., output variables.

[0040] By studying the mapping relationship between matrix interference factors and absorbance effects in a milk matrix interference sample set, a model capable of correcting absorbance deviations under actual matrix interference conditions is established. The obtained milk matrix interference sample set is divided into a training subset and a preliminary validation subset according to a predetermined ratio. The input variables for the training subset are matrix interference factor data, such as fat content, lactose concentration, non-target protein content, and pH value. The output variable is the corresponding milk absorbance effect data, i.e., the absorbance detection value under that interference condition. Subsequently, the training subset is input into a support vector machine (SVM). An appropriate kernel function, such as a radial basis function (RBF) kernel, is selected to handle the potential nonlinear correlation between interference factors and absorbance. A reasonable penalty coefficient is set to balance the model's fitting accuracy and generalization ability. Through iterative optimization of the SVM's decision boundary, the model can accurately predict the interference effect of the input matrix interference factor parameters on absorbance, thereby correcting the deviation of the absorbance detection value. During training, the model's correction error is evaluated in real time using the initial validation subset. When the error stabilizes within a preset range and the model converges, training is stopped, and an initial matrix interference corrector with absorbance deviation correction function is finally generated.

[0041] For the obtained initial matrix interference corrector, cross-validation was used. k-fold cross-validation evaluated its correction accuracy and generalization ability. The obtained milk matrix interference sample set was randomly divided into k subsets of similar size. One subset was selected as the validation set, and the remaining k-1 subsets were used as the training set. This training and validation process was repeated k times. Each time, the absorbance correction error of the initial corrector (mean square error, absolute error, etc.) was calculated based on the validation set. The average of the k results was used to determine the overall performance weakness of the initial corrector. Subsequently, based on the error sources identified by cross-validation, such as large correction bias under the gradient of specific matrix interference factors or insufficient adaptability of the model to certain milk matrix types, the core parameters of the support vector machine were adjusted, such as kernel function type, penalty coefficient, and gamma value. A small amount of targeted interference sample data was also added. If data was insufficient under a certain type of interference, the corrector was retrained. Through repeated iterative training and cross-validation, the correction error is continuously reduced and the adaptability of the corrector to different matrix interference conditions is improved until the average correction error of cross-validation is stably lower than the preset threshold, and the correction results of the corrector under the gradient combination of various interference factors can meet the accuracy requirements of milk A2 protein detection. At this point, the iteration is stopped, and a stable and accurate milk matrix interference corrector is finally established.

[0042] In one possible implementation, step S200 further includes:

[0043] Step S210: Obtain the ELISA technology detection process, extract nodes from the ELISA technology detection process, and obtain the ELISA detection node set.

[0044] Step S220: Analyze the detection steps of each detection node in the ELISA detection node set to generate an ELISA node detection program.

[0045] Step S230: Optimize the control strategy parameters based on the ELISA node detection program to construct an ELISA adaptive detection module.

[0046] Specifically, based on the technical requirements for milk A2 protein detection, a standard ELISA detection procedure suitable for this detection scenario is obtained. This procedure needs to cover the entire chain of operations from the interaction of milk samples and reagents to the reading of absorbance signals. Specifically, it includes the core operation steps such as adding the sample to be tested, incubating with A2 protein-specific primary antibody to achieve specific binding of antigen and antibody, rinsing with washing solution to remove unbound primary antibody, secondary antibody incubation, secondary washing to remove unbound secondary antibody, adding chromogenic reagent and reacting, and reading absorbance values ​​with an ELISA reader. Subsequently, based on the principles of independent operational purpose and indivisible process, the above complete ELISA detection process was divided into nodes: each operational step with independent function and which cannot be further subdivided was defined as a detection node. For example, adding the sample to be tested was defined as the sample addition node, primary antibody incubation was defined as the primary antibody incubation node, washing to remove unbound primary antibody was defined as the first washing node, secondary antibody incubation was defined as the secondary antibody incubation node, second washing to remove unbound secondary antibody was defined as the second washing node, adding chromogenic reagent and reaction was defined as the chromogenic reaction node, and reading absorbance by the microplate reader was defined as the absorbance detection node. Finally, all the defined independent nodes were integrated to form a structured ELISA detection node set.

[0047] For the obtained ELISA detection node set, the detection steps of each node were analyzed one by one. The core operational purpose of each node was clarified, the required reagent types and dosage ranges were determined, the operational timing requirements of each node were outlined, and the basic control parameter boundaries of each node were defined. Simultaneously, the start-up trigger conditions and termination criteria for node operations were clarified. Subsequently, this analyzed structured information was transformed into a programmed language recognizable by the detection equipment. This generated an independent program unit for each detection node, containing complete logic for parameter setting, operation execution, status judgment, and result feedback—the ELISA node detection program. This ensures that the operation of each node can be accurately controlled and traced, providing a programmatic foundation for the subsequent construction of the ELISA adaptive detection module.

[0048] For each generated ELISA node detection program, the impact of different control parameters on the detection results in each node program is analyzed, and the key control parameter set of the node program that plays a crucial role in detection accuracy and efficiency is screened out. Next, a set of milk sample types covering different types is obtained. For each type of milk sample in this set, the key control parameter set is analyzed to clarify the reasonable value range of each key control parameter under different milk types, forming the milk type-node control strategy parameter threshold. Subsequently, an ELISA milk protein detection target is preset. This target clarifies the requirements for accuracy, time, etc., during the detection process. Based on this target, the control strategy parameters are optimized within the milk type-node control strategy parameter threshold to determine the optimal node control strategy parameters suitable for different milk types, i.e., the milk type-target node control strategy parameters. Finally, the milk type-target node control strategy parameters are identified and stored, and integrated into the milk protein detection module. This allows the module to automatically match the corresponding optimal control strategy parameters according to the type of milk sample to be detected, ultimately constructing an ELISA adaptive detection module with adaptive adjustment capabilities.

[0049] In one possible implementation, step S230 further includes:

[0050] Step S231: Perform a control parameter impact analysis on each node program in the ELISA node detection program, and screen the key control parameter set of the node program.

[0051] Step S232: Obtain the milk sample type set, and perform strategy parsing on the key control parameter set of the node program according to the milk sample type set to obtain the milk type-node control strategy parameter threshold.

[0052] Step S233: Preset the ELISA milk protein detection target, optimize the control strategy parameters within the threshold of the milk type-node control strategy parameters based on the ELISA milk protein detection target, and construct the ELISA adaptive detection module.

[0053] Specifically, the composition of the ELISA detection procedure is clearly defined. This procedure includes multiple independent steps such as sample loading, primary antibody incubation, washing, secondary antibody incubation, colorimetric reaction, and absorbance detection. Each step corresponds to a series of adjustable control parameters, such as the loading volume and loading rate for the sample loading step, the incubation temperature and incubation time for the primary antibody incubation step, the number of washes and the wash buffer flow rate for the washing step, and the amount of chromogenic reagent and reaction time for the colorimetric reaction step. Subsequently, an influence analysis of the control parameters is conducted for each step: by using the controlled variable method, keeping other parameters constant and only changing the value of a certain target parameter, the output results of the step procedure under different values ​​are compared. For example, the correlation between binding efficiency, impurity residue, absorbance stability, etc., and overall detection performance, such as detection accuracy and repeatability, is analyzed to determine the degree of influence of the parameter on the detection results, such as strong influence, weak influence, or no influence. Finally, based on the impact analysis results, control parameters that play a decisive role in the implementation of each node program function and the overall detection performance are selected, such as incubation temperature, incubation time, and number of washes. These parameters are then categorized and integrated according to the node program to form a set of key control parameters for the node program. This ensures that subsequent parameter optimization can focus on core variables, thereby improving optimization efficiency and the performance of the detection module.

[0054] Based on the common differences in milk types encountered in actual testing scenarios, a set of milk sample types covering different matrix characteristics is obtained. This set typically includes representative milk types such as whole milk, low-fat milk, skim milk, pasteurized milk, and UHT milk. These types differ significantly in terms of fat content, protein denaturation degree, and impurity composition, which will have different impacts on the ELISA detection process. Next, based on the selected set of key control parameters for each node procedure, such as incubation temperature, incubation time, number of washes, and amount of chromogenic reagent, control parameter strategies are analyzed for each type of milk in the sample type set. For a specific milk type, the detection effect under different values ​​of key control parameters is tested through multiple pre-experiments, such as absorbance stability and antigen-antibody binding efficiency. The matrix characteristics of this milk type are analyzed; for example, the high fat content of whole milk may require a longer incubation time to ensure adequate binding and adaptation to the requirements of each key control parameter. This allows for the determination of reasonable value ranges for key control parameters at each node procedure for this milk type, i.e., the upper and lower limits of the parameters. Finally, the node programs, key control parameters, and value ranges corresponding to all milk types are linked and integrated to form a structured milk type-node control strategy parameter threshold.

[0055] Based on the actual needs and accuracy requirements of milk A2 protein detection, specific ELISA milk protein detection targets are preset. These targets typically encompass detection accuracy (e.g., absorbance measurement error must be below a specific value), protein content calculation deviation must be controlled within a preset range), detection efficiency (e.g., the total time for a single detection must not exceed a specified time), and detection stability (e.g., the coefficient of variation of multiple parallel detection results must meet standards). These core dimensions provide clear guidance for subsequent parameter optimization. Next, using the obtained milk type-node control strategy parameter thresholds as constraints, for each type of milk sample, within the corresponding key control parameter thresholds for each node procedure, the control strategy parameters are optimized through a combination of parameter combination testing and performance evaluation. For example, for whole milk samples, within the temperature threshold (e.g., 36℃~38℃) and time threshold (e.g., 30min~40min) of the primary antibody incubation node, detection results under different temperature and time combinations are tested to screen out the optimal parameter combination that meets the preset detection targets and is suitable for the characteristics of the whole milk matrix. Finally, the node program and optimal control strategy parameters corresponding to all milk types are associated and identified, and stored in the milk protein detection module. This enables the module to automatically identify the sample type and match the corresponding optimal control strategy parameters when detecting different types of milk samples, thus constructing an ELISA adaptive detection module with adaptive adjustment capabilities.

[0056] In one possible implementation, step S233 further includes:

[0057] Step S2331: Construct a milk protein detection target function based on the ELISA milk protein detection target.

[0058] Step S2332: Optimize the control strategy parameters within the threshold of the milk type-node control strategy parameters using the milk protein detection objective function to determine the milk type-target node control strategy parameters.

[0059] Step S2333: Store the milk type-target node control strategy parameter identifier into the milk protein detection module to construct the ELISA adaptive detection module.

[0060] Specifically, defining the core dimensions of ELISA milk protein detection targets typically includes detection accuracy (A2 protein content calculation deviation ≤5%), detection efficiency (single test duration ≤60 minutes), and detection stability (e.g., parallel sample absorbance variation coefficient ≤3%). These dimensions need to be determined based on actual testing needs and industry standards. Next, each target dimension is quantified: detection accuracy is converted into a minimum absorbance measurement error index, detection efficiency into a minimum total test time index, and detection stability into a minimum result variation coefficient index. Simultaneously, different weights are assigned according to the importance of each dimension, such as detection accuracy weighted at 0.5, detection efficiency weighted at 0.3, and detection stability weighted at 0.2, ensuring that key targets are prioritized. Finally, using key control parameters of each node program, such as incubation temperature, time, and number of washes, as variables, the quantified indicators and weights are integrated to construct a milk protein detection objective function in the form of a mathematical expression. The function calculates the comprehensive target value in the form of (absorbance error × 0.5) + (detection time × 0.3 / 60) + (coefficient of variation × 0.2 / 3). Multi-objective collaborative optimization is achieved by minimizing this comprehensive value.

[0061] The optimization process employs a dual constraint: a constructed milk protein detection objective function that comprehensively reflects the requirements for detection accuracy, efficiency, and stability; and threshold values ​​for milk type-node control strategy parameters as boundary constraints. This ensures the optimization process aligns with both the detection objective and practical feasibility. Next, for each milk type in the sample set, such as whole milk and skim milk, parameter optimization is performed individually. Key control parameters for each node program corresponding to this milk type, such as incubation temperature, incubation time, and number of washes, are used as variables. Multiple combinations are tested within their corresponding threshold values, and each parameter combination is substituted into the milk protein detection objective function to calculate the comprehensive target value. By comparing the comprehensive target values ​​of different combinations, the parameter combination that optimizes the objective function value is selected. A minimum comprehensive target value corresponds to low detection error, high efficiency, and good stability. Finally, the optimal parameter combination for each milk type is correlated with the milk type and node program to determine the milk type-target node control strategy parameters. This ensures that different milk types can be matched with suitable optimal control parameters in subsequent detections.

[0062] The defined milk type-target node control strategy parameters are structurally labeled, clearly indicating the corresponding milk type (e.g., whole milk, skim milk), the associated detection node (e.g., primary antibody incubation, washing, colorimetric reaction), and specific parameter values ​​for each parameter combination (e.g., primary antibody incubation: temperature 37℃, time 35 min; washing: 5 times, flow rate 2 mL / s). This ensures a clear and traceable correspondence between parameters and application scenarios. Next, these labeled parameters are imported into the parameter database of the milk protein detection module, establishing a mapping mechanism between milk type, detection node, and optimal control parameters. This enables the module to automatically retrieve and call the optimal parameters for the corresponding node based on the milk type. Finally, the module's parameter calling logic is debugged and verified to ensure that when the type of milk sample to be tested is input, the module can quickly match the corresponding target node control strategy parameters and drive each detection node to perform operations according to the optimal parameters. This ultimately constructs an adaptive ELISA detection module that can adapt to different milk types.

[0063] In one possible implementation, step S300 further includes:

[0064] Step S310: Construct a protein detection pretreatment program, which includes low-temperature defatting, protein denaturation inhibition, and multi-stage filtration.

[0065] Step S320: Obtain the target milk sample and perform detection preprocessing on the target milk sample according to the protein detection preprocessing procedure to obtain a usable milk sample.

[0066] Step S330: Homogenize and aliquot the available milk sample to obtain the milk sample to be tested.

[0067] Specifically, the design basis of the protein detection pretreatment procedure was clarified. Combining the characteristics of the milk matrix and the ELISA detection principle, and targeting interfering factors in milk such as fat, denatured proteins, and suspended impurities that may affect the detection of A2 protein, the pretreatment procedure was determined to include three core operational modules. The low-temperature defatting module separates the fat components in milk from the liquid phase by setting a specific low-temperature environment, such as 4℃-8℃, and using methods such as centrifugation or static separation. This prevents fat particles from hindering the specific binding of antigen and antibody or interfering with absorbance signal reading in subsequent detection. The protein denaturation inhibition module inhibits the activity of endogenous proteases in milk by adding appropriate protease inhibitors, such as benzyl sulfonyl fluoride, or by strictly controlling the processing temperature, preventing structural denaturation of the target A2 protein and ensuring that it retains its natural antigen activity for effective antibody binding. The multi-stage filtration module uses filter membranes of different pore sizes. For example, a 0.45μm filter membrane is first used to remove larger particulate impurities, followed by a 0.22μm filter membrane to filter out small suspended matter and large molecular interfering substances, further purifying the sample matrix through a stepwise filtration process. Finally, these three major operation modules are integrated in the logical order of low-temperature defatting, protein denaturation inhibition, and multi-stage filtration to form a protein detection pretreatment procedure with a clear process and well-defined parameters, ensuring that the pretreated milk samples can meet the requirements of ELISA detection for matrix purity and target protein activity.

[0068] Obtain target milk samples according to actual testing needs. Sample types can cover common categories such as raw milk, pasteurized milk, UHT milk, whole milk, low-fat milk, and skim milk to ensure that the samples are representative for testing. At the same time, basic information of the samples, such as production batch and storage conditions, should be recorded to provide a basis for subsequent traceability and analysis of test results. Next, strictly following the constructed protein detection pretreatment procedure, the target milk sample was processed in stages: The first stage involved low-temperature defatting, where the milk sample was placed in a preset low-temperature environment (usually 4℃~8℃) for a period of time, followed by centrifugation (e.g., 3000-5000 rpm for 10-15 minutes) to separate and remove the fat layer, reducing interference from fat particles on subsequent antigen-antibody binding and absorbance detection. The second stage involved protein denaturation inhibition, where protease inhibitors were added to the defatted milk sample in a specific ratio and gently mixed to inhibit endogenous protease activity, preventing structural denaturation of the target A2 protein and ensuring its antigenic activity. The third stage involved multi-stage filtration, where the sample, after the first two steps, was sequentially passed through filter membranes of different pore sizes. For example, a 0.45μm microporous membrane was used to remove larger suspended impurities, followed by a 0.22μm microporous membrane to remove small particles and large interfering molecules, completing sample purification. After this complete pretreatment process, a usable milk sample with a pure matrix and stable A2 protein activity was obtained.

[0069] The obtained usable milk samples are transferred to a homogenization device, and appropriate homogenization parameters are set. For example, the speed of a high-speed dispersion homogenizer can be set to 8000~12000 rpm, and the processing time can be 3~5 minutes; or the power of an ultrasonic homogenizer can be set to 200~300W, and the processing time can be 2~4 minutes. The mechanical or ultrasonic action of the device breaks up any small aggregates that may exist in the sample, so that the components in the sample, including the target A2 protein, are evenly distributed, avoiding deviations in results between different detection wells or different batches due to uneven sample concentration. After homogenization, the milk sample is aliquoted and stored as follows: Based on the sample volume required for a single ELISA test, such as 50-100 μL per well, the homogenized milk sample is accurately aliquoted into sterile centrifuge tubes or disposable test tubes using a sterile pipette. The aliquot volume in each tube should be slightly more than the volume required for a single test to allow for operational margin. After aliquoting, the test tubes are sealed, such as by adding a sterile sealing cap or applying a sealing film. Depending on the short-term or long-term storage requirements of the sample, it is stored in the corresponding environment to obtain a stable milk sample that can be directly used for subsequent testing.

[0070] In one possible implementation, step S300 further includes:

[0071] Step S340: Identify the target milk type of the milk sample to be tested.

[0072] Step S350: Control the ELISA dynamic detection module to perform control strategy parameter matching and protein optimization detection on the target milk type to obtain the milk absorbance detection data.

[0073] Specifically, the identification dimensions of the milk sample type to be tested are clearly defined, combining common milk classification standards in actual testing scenarios. These mainly cover core dimensions such as fat content (e.g., whole milk, low-fat milk, skim milk), sterilization method (e.g., pasteurized milk, UHT milk), and processing technology (e.g., fresh milk, modified milk). This ensures that the identification results cover the type differences required for subsequent parameter matching. Next, identification is carried out using information verification and feature-assisted verification: on the one hand, the original information of the milk sample to be tested is prioritized, such as production labels and testing records, extracting clear information such as sample type, fat content, and sterilization method from the label as preliminary identification basis; on the other hand, if the sample information is incomplete or ambiguous, rapid feature detection can be used to assist in judgment, such as detecting sample fat content to distinguish between whole milk and skim milk, and indirectly verifying the sterilization method by detecting microbial residues, reducing identification errors. Finally, the identification results are classified and labeled according to a unified standard, such as whole milk pasteurized milk or skim UHT milk, to determine the target milk type of the milk sample to be tested.

[0074] Based on the identified target milk type, the parameter matching function of the ELISA dynamic detection module is activated. This module has a built-in database of milk type-target node control strategy parameters, which automatically retrieves and calls up the optimal control strategy parameters corresponding to the current target milk type. This includes key control parameters for each detection node, such as sample loading, primary antibody incubation, washing, and colorimetric reaction, such as incubation temperature, incubation time, number of washes, and amount of colorimetric reagent, ensuring a high degree of fit between the detection parameters and the sample type. Next, under the control of the matched control strategy parameters, the module performs protein optimization detection on the milk sample to be tested: first, the milk matrix interference corrector in the module, combined with the previously trained interference correction model, eliminates the interference of residual fat, impurities, and other factors in the sample matrix on absorbance detection in real time; then, following the ELISA detection procedure, it sequentially performs operations such as sample loading, primary antibody specific binding, multiple rounds of washing to remove unbound substances, secondary antibody incubation, and colorimetric reaction. Each step is performed under the control of the appropriate parameters, ensuring the stability and accuracy of the detection process. Finally, the module reads the absorbance signal after the colorimetric reaction in real time through the built-in microplate reader, performs preliminary filtering and calibration on the signal, and finally generates milk absorbance detection data that can be used for subsequent protein content calculation.

[0075] In one possible implementation, step S400 further includes:

[0076] Step S410: Prepare a standard A2 protein milk sample, and use ELISA technology to detect and read the standard A2 protein milk sample to obtain standard milk A2 protein concentration data and standard milk absorbance data.

[0077] Step S420: Based on the standard milk A2 protein concentration data and standard milk absorbance data, perform logistic regression fitting to establish the absorbance-concentration standard curve.

[0078] Specifically, the preparation of standard A2 protein milk samples was carried out. High-purity A2 protein standards were selected as the core raw material, and base milk that was free of A2 protein and had stable matrix composition was selected as the dilution carrier. According to the concentration range requirements of subsequent detection, at least five gradient standard concentrations were designed, such as 0.05 μg / mL, 0.2 μg / mL, 1 μg / mL, 5 μg / mL, and 20 μg / mL. The A2 protein standards were added to the base milk step by step according to the calculated amount using a precision pipetting device. Each addition was thoroughly mixed to ensure that the standards were uniformly dissolved in the milk matrix and to avoid uneven concentration distribution. Finally, a series of standard A2 protein milk samples with known concentrations and clear gradients were obtained. Subsequently, the prepared standard samples were analyzed and read using ELISA technology. Following the standard ELISA procedure, the following steps were performed: adding standard samples sequentially, setting up three parallel samples for each concentration gradient to reduce errors; primary antibody incubation, reacting at the appropriate temperature for a preset time to ensure specific binding of A2 protein to the primary antibody; washing to remove unbound primary antibody and impurities; secondary antibody incubation, adding enzyme-labeled secondary antibody to react with the bound primary antibody; secondary washing; color development reaction; adding substrate solution; and enzyme catalysis to produce color change. After the color development reaction reached a stable state, the absorbance value of each standard sample well was read using an ELISA reader at a specific wavelength, usually 450 nm, adjusted according to the type of chromogenic substrate. The absorbance data of each parallel sample was recorded and the average value was calculated. At the same time, the known A2 protein concentrations of each standard sample were compiled, ultimately forming a dataset with a one-to-one correspondence between standard milk A2 protein concentration data and standard milk absorbance data.

[0079] Using the obtained standard milk A2 protein concentration data as the x-axis and the corresponding standard milk absorbance data as the y-axis, the data are imported into data processing software such as Excel or Origin. A logistic regression algorithm is used to fit and analyze these data points, constructing a mathematical model that reflects the quantitative relationship between absorbance and A2 protein concentration. The model must meet the linear correlation requirement within the detection range, such as a correlation coefficient R² ≥ 0.99. The fitted curve is then visualized, and its corresponding regression equation is recorded, such as Y = aX + b, where a and b are the fitting coefficients. Finally, an absorbance-concentration standard curve is established that can be used for subsequent data conversion.

[0080] In one possible implementation, step S400 further includes:

[0081] Step S430: Based on the absorbance-concentration standard curve, convert the milk absorbance detection data to protein content to obtain initial milk A2 protein content detection data.

[0082] Step S440: Perform spiked recovery testing and result correction on the initial milk A2 protein content detection data to determine the milk A2 protein content detection result.

[0083] Specifically, the conversion is based on an established absorbance-concentration standard curve. This curve, fitted using logistic regression, accurately reflects the quantitative relationship between the absorbance value and corresponding concentration of a standard A2 protein milk sample, and includes a fitting equation that can be directly used for calculation. Next, the acquired milk absorbance detection data is processed. This data has been processed using the ELISA dynamic detection module to eliminate milk matrix interference, ensuring that the absorbance value is only related to the actual A2 protein content in the sample. Subsequently, the absorbance value of each milk sample to be tested is substituted into the fitting equation of the absorbance-concentration standard curve, and the corresponding A2 protein concentration value is calculated through inverse equation calculation. If the milk sample to be tested was diluted during the pretreatment stage, such as a gradient dilution to adapt to the linear detection range, the calculated concentration value needs to be converted according to the actual dilution factor. For example, if diluted 10 times, the calculated concentration is multiplied by 10. Finally, preliminary data on the A2 protein content per unit volume of the milk sample to be tested is obtained, i.e., the initial milk A2 protein content detection data.

[0084] Spiked recovery testing was conducted by selecting a portion of the milk samples to be tested as the spiking targets. A known concentration of A2 protein standard was precisely added to these samples, ensuring the spiking amount remained within the linear range of the detection method and avoided exceeding the upper limit of the standard curve concentration. Subsequently, the spiked samples were re-tested following the same pretreatment procedures as the samples to be tested, such as low-temperature defatting and multi-stage filtration, as well as the ELISA detection procedure, including parameter matching and absorbance reading. The A2 protein content was calculated after spiking. The recovery rate was calculated using the formula: (Spiked recovery value - Initial recovery value) / Spiked amount × 100%. This verified the accuracy of the detection method. The recovery rate was required to be within a reasonable range of 80% to 120%. If the recovery rate was within this range, it indicated that the detection method had no significant systematic error and the initial detection data was reliable. Next, the results are corrected. Taking into account potential error factors in the previous detection process, such as the residual interference coefficient of the milk matrix interference corrector, the small deviation of the standard curve fitting, and the system error calibration value of the detection instrument, the initial milk A2 protein content detection data are fine-tuned. For example, if the matrix interference corrector has an average residual error of 0.5%, then an error compensation of 0.5% is required based on the initial data to finally obtain the milk A2 protein content detection result that has been verified and corrected and meets the detection requirements.

[0085] It should be noted that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, the above description focuses on specific embodiments of this specification. Additionally, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some implementations, multitasking and parallel processing are possible or may be advantageous.

[0086] The above description is only a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

[0087] This specification and accompanying drawings are merely illustrative examples of this application and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Clearly, those skilled in the art can make various alterations and modifications to this application without departing from its scope. Therefore, if such modifications and variations fall within the scope of this application and its equivalents, this application intends to include such modifications and variations.

Claims

1. A method for optimizing the detection of the A2 protein content in milk based on the ELISA technique, characterized by, The method comprises: Collecting a milk matrix interference detection data set, performing correction training fitting based on the milk matrix interference detection data set, and establishing a milk matrix interference corrector; Constructing an ELISA adaptive detection module, coupling the milk matrix interference corrector to the ELISA adaptive detection module, and generating an ELISA dynamic detection module; Preparing a milk sample to be detected, controlling the ELISA dynamic detection module to perform protein optimization detection on the milk sample to be detected, and obtaining milk absorbance detection data; Establishing an absorbance-concentration standard curve, performing protein content conversion on the milk absorbance detection data based on the absorbance-concentration standard curve, and determining a milk A2 protein content detection result; Collecting a milk matrix interference detection data set, comprising: According to the milk A2 protein detection technology, a set of milk matrix interference factors is obtained; Gradient analysis is performed on each interference factor in the set of milk matrix interference factors to determine a set of matrix interference factor gradient levels; Based on the set of matrix interference factor gradient levels, an orthogonal test design is performed to construct a matrix interference factor test parameter table; Using ELISA technology, milk matrix interference testing is performed based on the matrix interference factor test parameter table to collect the milk matrix interference detection data set; Establishing a milk matrix interference corrector, comprising: According to the milk matrix interference detection data set, a set of matrix interference factor data and a corresponding set of milk absorbance influence data are obtained; The set of matrix interference factor data is used as an input variable, the corresponding set of milk absorbance influence data is used as an output variable, the input variable and the output variable are divided and identified to obtain a set of milk matrix interference samples; Using a support vector machine, absorbance correction training fitting is performed on the set of milk matrix interference samples to obtain an initial matrix interference corrector; The initial matrix interference corrector is cross-validated and iteratively optimized to establish the milk matrix interference corrector; Constructing an ELISA adaptive detection module, comprising: Obtaining an ELISA technology detection process, extracting nodes from the ELISA technology detection process to obtain a set of ELISA detection nodes; Performing detection step analysis on each detection node in the set of ELISA detection nodes to generate an ELISA node detection program; Based on the ELISA node detection program, control strategy parameter optimization is performed to construct an ELISA adaptive detection module; Based on the ELISA node detection program, control strategy parameter optimization is performed to construct an ELISA adaptive detection module, comprising: Performing control parameter influence analysis on each node program in the ELISA node detection program to screen a set of node program key control parameters; Obtaining a set of milk sample types, and performing strategy analysis on the set of node program key control parameters according to the set of milk sample types to obtain a milk type-node control strategy parameter threshold; Presetting an ELISA milk protein detection target, performing control strategy parameter optimization within the milk type-node control strategy parameter threshold based on the ELISA milk protein detection target to construct an ELISA adaptive detection module; Based on the ELISA milk protein detection target, the control strategy parameter optimization is performed within the milk type-node control strategy parameter threshold value, and an ELISA adaptive detection module is constructed, including: According to the ELISA milk protein detection target, a milk protein detection target function is constructed; Using the milk protein detection target function, the control strategy parameter optimization is performed within the milk type-node control strategy parameter threshold value, and the milk type-target node control strategy parameter is determined; The milk type-target node control strategy parameter is identified and stored in the milk protein detection module to construct the ELISA adaptive detection module.

2. The method for optimizing the detection of A2 protein content in milk based on ELISA technology as claimed in claim 1, wherein, Prepare the milk sample to be detected, including: Construct a protein detection pretreatment program, which includes low-temperature degreasing, protein denaturation inhibition, and multi-stage filtration; Obtain the target milk sample, and detect and pretreat the target milk sample according to the protein detection pretreatment program to obtain a usable milk sample; Homogenize and store the usable milk sample to obtain the milk sample to be detected.

3. The method for optimizing the detection of A2 protein content in milk based on ELISA technology as claimed in claim 1, wherein, Get the milk absorbance detection data, including: Identify the target milk type of the milk sample to be detected; Control the ELISA dynamic detection module to perform control strategy parameter matching and protein optimization detection on the target milk type to obtain the milk absorbance detection data.

4. The method for optimizing the detection of A2 protein content in milk based on ELISA technology as claimed in claim 1, wherein, Establish the absorbance-concentration standard curve, including: Prepare a standard A2 protein milk sample, detect and read the standard A2 protein milk sample using ELISA technology, and obtain standard milk A2 protein concentration data and standard milk absorbance data; Based on the standard milk A2 protein concentration data and standard milk absorbance data, perform logistic regression fitting to establish the absorbance-concentration standard curve.

5. The method for optimizing the detection of A2 protein content in milk based on ELISA technology as claimed in claim 1, wherein, Determine the milk A2 protein content detection result, including: Based on the absorbance-concentration standard curve, the protein content of the milk absorbance detection data is converted to obtain the initial milk A2 protein content detection data; The initial milk A2 protein content detection data is subjected to standard addition recovery test and result correction to determine the milk A2 protein content detection result.

Images