Multi-quantitative detection method for cytokines

Abstract

The invention provides a multiple quantitative detection method for cytokines, and belongs to the technical field of cytokine detection.The multiple quantitative detection method comprises the steps that firstly, a sample is subjected to centrifugal separation and dilution pretreatment, and various data parameters in the detection process are collected; a standard product is prepared and calibrated through a standard freeze-drying process, and a concentration gradient sequence is established. And constructing a calibration model by adopting a multivariate mixed decomposition method, and determining a detection range. The detection conditions are optimized through an orthogonal test, and the function relationship among the parameters is established. In order to eliminate the difference between batches, a correction model is constructed and standardized processing is carried out. And establishing a three-level quality control system to monitor the stability of the system, and developing a signal cross interference elimination model to perform error compensation. Finally, standardization processing is conducted on detection data, a characteristic fusion algorithm is adopted for correcting signals, accurate quantitative detection of the cytokines is achieved in combination with a calibration curve and a quality control evaluation system, and the problem that the accuracy of the detection result is not high in an existing cytokine multiple quantitative detection technology is solved.

Description

Technical Field

[0001] This invention belongs to the field of cytokine detection technology, and more specifically, relates to a method for multiplex quantitative detection of cytokines. Background Technology

[0002] Cytokines, as important immune inflammatory factors, play a crucial role in disease diagnosis, inflammatory responses, immune responses, and treatment monitoring. With the development of precision medicine, the demand for multiplex quantitative detection of cytokines in clinical practice is increasing. Currently, commonly used clinical methods for cytokine detection mainly include enzyme-linked immunosorbent assay (ELISA), flow cytometry, and multiplex bead array technology. While traditional ELISA technology has high sensitivity and specificity, it can only detect a single cytokine per assay, resulting in low throughput, high sample consumption, and cumbersome and time-consuming operation. Flow cytometry can simultaneously detect multiple parameters, but the equipment is expensive, requires highly skilled operators, and the interpretation of results is somewhat subjective. Multiplex bead array technology, by conjugating microspheres with different fluorescent labels to specific antibodies, can simultaneously detect multiple cytokines, significantly improving detection efficiency.

[0003] However, existing multiplex quantification techniques for cytokines still have the following technical problems: First, batch-to-batch differences during the detection process lead to poor comparability and reproducibility of the results; second, cross-interference occurs when multiple cytokines are detected simultaneously, affecting the accuracy of the detection; and third, the lack of a systematic quality control system makes it difficult to guarantee the reliability of the results.

[0004] In summary, existing multiplex quantification techniques for cytokines suffer from low accuracy. Summary of the Invention

[0005] In view of this, the present invention provides a method for multiplex quantification of cytokines, which can solve the problem of low accuracy of detection results in existing multiplex quantification techniques for cytokines.

[0006] This invention is implemented as follows: It provides a method for multiplex quantitative detection of cytokines, comprising the following steps: standardizing the preprocessing of cytokine detection samples; real-time acquisition of fluorescence signal data, background signal data, standard concentration data, sample dilution rate, reaction temperature, and ambient humidity during the detection process; preparing standard sequences and using a standard lyophilization process to prepare cytokine standards; constructing a calibration model, constructing the standard data into a sparse matrix and performing multivariate mixture decomposition to obtain stable and variable signal components, and calculating the calibration coefficient matrix based on the decomposition results; optimizing the multiplex detection conditions of cytokines using orthogonal experiments; establishing a batch-to-batch difference correction model, constructing a batch-to-batch difference sparse matrix and performing multivariate mixture decomposition to obtain stable and variable components; designing a three-level quality control system, preparing high-concentration, medium-concentration, and low-concentration quality control products; establishing a signal cross-interference elimination model, constructing a cross-interference feature sparse matrix and performing multivariate mixture decomposition; standardizing the detection data; and establishing a detection result evaluation system to calculate the final concentration of cytokines in the sample.

[0007] The standardized preprocessing of cytokine detection samples includes: collecting and centrifuging the samples, determining the centrifugation speed to be 500 to 5000 rpm and the centrifugation time to be 5 to 30 minutes depending on the sample type; serially diluting the supernatant after separation, with dilution ratios ranging from 1:2 to 1:100; using a standardized acquisition system to record fluorescence signal data in real time, including target signal intensity, background signal intensity, and internal reference signal intensity; synchronously acquiring the detection system's operating parameters; and establishing a data preprocessing workflow to perform baseline drift correction, signal smoothing, and peak identification on the acquired signals.

[0008] The preparation of standard sequences includes: preparing stock solutions of cytokine standards; dissolving and lyophilizing the standards in ultrapure water; converting the concentrations according to molecular weight; preparing cytokine standards using a standard lyophilization process, controlling the temperature gradient, vacuum level, and amount of protective agent added during the lyophilization process; calibrating using international standards, calculating the relative calibration factor, and establishing the conversion relationship between standard concentrations and international units; preparing a standard dilution sequence with an 8-point concentration gradient, setting the dilution factor between adjacent concentration points to 2 times to cover the linear range of the detection system; and converting the standard sequence data into a sparse matrix form.

[0009] The calibration model construction includes: constructing a sparse matrix from the standard data, where the number of rows in the sparse matrix represents the number of standard concentration gradients and the number of columns represents the number of cytokine types; performing multi-scale decomposition on the sparse matrix, extracting different frequency components using wavelet transform, and selecting the optimal decomposition level; adaptively decomposing the signal using empirical mode decomposition, extracting intrinsic mode functions, and reconstructing and optimizing them; separating stable and variable components of the signal using independent component analysis, and establishing a mathematical model of the signal components; constructing a reconstruction optimization objective function, and optimizing the reconstruction parameters using an iterative algorithm.

[0010] The optimization of multiplex cytokine detection conditions using orthogonal experiments included: designing an orthogonal experimental scheme, selecting a reaction temperature range of 20 to 40 degrees Celsius, an incubation time range of 1 to 24 hours, and an oscillation frequency range of 200 to 800 rpm; calculating the contribution of each factor to the detection results, and using analysis of variance to determine the main effect and interaction effect; establishing a factor weight calculation model based on the entropy weight method to determine the weight coefficients of each parameter; constructing a parameter optimization objective function, and using nonlinear programming to solve for the optimal parameter combination; and verifying the stability and reproducibility of the optimized parameter combination.

[0011] The process of establishing a batch-to-batch variance correction model includes: constructing a sparse matrix of batch-to-batch variances and calculating the within-batch and between-batch variances; performing multivariate decomposition on the variance matrix and using principal component analysis to extract the main sources of variation; calculating the variance decomposition ratio and establishing a weight determination model based on the entropy method; introducing the Fisher information matrix to evaluate parameter sensitivity and optimize the correction coefficients; and using cross-validation to evaluate the stability of the correction model.

[0012] The design of the three-level quality control system includes: preparing quality control samples at three concentration levels (high, medium, and low) and selecting typical concentration ranges of the target analyte; establishing a local feature extraction model, including signal peak shape features, baseline features, and noise features; constructing a global feature analysis model, including system stability, linear range, and reproducibility indicators; designing a feature fusion algorithm that integrates local and global features using a weighted summation method; and establishing a quality control early warning mechanism, setting early warning thresholds and alarm rules.

[0013] The establishment of a signal cross-interference elimination model includes: constructing a sparse matrix of cross-interference features and establishing an interaction strength function; using a competitive inhibition model to describe the interaction relationship between cytokines; calculating the steady-state solution to obtain the cross-interference matrix and optimizing the nonlinear correction term; using an alternating minimization algorithm to optimize the correction parameters; and establishing an iterative optimization program to achieve dynamic parameter adjustment.

[0014] The standardization process for the detection data includes: converting the original detection signal into a sparse matrix form and establishing a signal processing model; using a block processing strategy to perform efficient calculations on large-scale data; designing a feature fusion algorithm to integrate multi-dimensional detection signals; establishing a correction parameter optimization model to achieve adaptive signal correction; generating the final quantitative results and evaluating the accuracy of the results.

[0015] The establishment of a testing result evaluation system includes: establishing a quality control product evaluation system and setting judgment standards for various indicators; using the analytic hierarchy process (AHP) to determine the weights of evaluation indicators; constructing a simulated annealing algorithm to optimize the weight coefficients; calculating the precision and accuracy of the system's testing results; and establishing a result reporting and quality traceability system.

[0016] Specifically, it includes the following steps:

[0017] S10. Standardize the pretreatment of cytokine detection samples, use centrifugation technology and determine the centrifugation speed and time according to the sample type, dilute the supernatant after separation according to the specified ratio, and collect fluorescence signal data, background signal data, standard concentration data, sample dilution rate and detection reaction temperature in real time during the detection process.

[0018] S20. Prepare standard sequence: Prepare cytokine standards using standard freeze-drying process and calibrate using international standards. Establish standard concentration gradient dilution sequence and organize standard concentration data and detection signal data into sparse matrix form.

[0019] S30. Construct a calibration model, construct the standard data into a sparse matrix form and perform multivariate hybrid decomposition to obtain the stable and variable components of the signal, calculate the calibration coefficient matrix based on the decomposition results, establish an 8-point concentration gradient calibration curve using the block matrix processing method, and determine the detection linear range and the minimum detection limit, including multivariate hybrid decomposition of the background signal data to obtain the stable and variable components of the background signal.

[0020] S40. Orthogonal experiments were used to optimize the multiple detection conditions for cytokines. Based on contribution analysis, the weighting coefficients of the incubation temperature, incubation time, and oscillation frequency of the reaction system were determined to achieve the optimal combination of detection parameters. The functional relationship between the detection reaction temperature, sample dilution rate, fluorescence signal variation component, and background signal variation component was established.

[0021] S50. Establish a batch-to-batch difference correction model, construct a batch-to-batch difference sparse matrix and perform multivariate hybrid decomposition to obtain stable and variable components, calculate the correction coefficient matrix based on matrix block technology, establish fluorescence signal variation coefficient matrix and background signal variation coefficient matrix according to the variation relationship, and standardize the detection results.

[0022] S60. Design a three-level quality control system, prepare high-concentration, medium-concentration, and low-concentration quality control products, calculate the fluorescence signal stability coefficient matrix and background signal stability coefficient matrix based on the fluorescence signal variation coefficient matrix and the background signal variation coefficient matrix, and use the local-global feature analysis method to monitor the system stability during the detection process.

[0023] S70. Establish a signal cross-interference elimination model, construct a cross-interference feature sparse matrix and perform multivariate hybrid decomposition, calculate the fluorescence signal error compensation amount and background signal error compensation amount based on the fluorescence signal stability coefficient matrix and the background signal stability coefficient matrix, calculate the cross-interference coefficient matrix based on the decomposition results, and use matrix operations to correct the original signal to obtain the cross-interference correction factor.

[0024] S80. Standardize the detection data, construct the original detection signal into a sparse matrix and divide it into blocks, apply the calculated fluorescence signal error compensation amount and background signal error compensation amount to the original fluorescence signal data and background signal data, and use the feature fusion algorithm to correct and combine each block signal to generate cytokine concentration quantification results.

[0025] S90. Establish a test result evaluation system, use corrected fluorescence signal data and background signal data combined with standard concentration data to establish a calibration curve, evaluate the test results of quality control products based on local-global feature analysis method, evaluate the precision and accuracy of test data according to feature contribution, and calculate the final concentration of cytokines in the sample.

[0026] The equations or calculation steps involved in this invention are described in detail below:

[0027] The process of constructing the standard curve in S30:

[0028] 1. First, construct the original data matrix M of the standard products. raw :

[0029]

[0030] In the formula, f ij denoted as the raw fluorescence signal value of the j-th cytokine at the i-th standard concentration point.

[0031] 2. Perform multivariate hybrid decomposition on the original data:

[0032] M raw =M stable +M var +M noise ;

[0033] The specific decomposition process is as follows:

[0034]

[0035]

[0036] In the formula, s ij To stabilize the signal components; v ij For varying signal components; n ij This represents the noise component.

[0037] 3. Establish the calibration curve equation:

[0038]

[0039] In the formula, C ij denoted as , where is the standard concentration of the j-th cytokine at the i-th concentration point; k1, k2, k3 are the fitting coefficients; b is the intercept; and ∈ represents the fitting error term.

[0040] Batch-to-batch difference correction in S50:

[0041] 1. Construct the batch difference matrix D:

[0042]

[0043] In the formula, d ij q represents the difference between the i-th batch and the reference batch in the j-th test indicator; q is the batch number; p is the number of test indicators.

[0044] 2. Calculate the correction coefficient matrix K:

[0045] K = D × W × T × H × R;

[0046] The weight matrix W is:

[0047]

[0048] The temperature effect matrix T is:

[0049]

[0050] The humidity effect matrix H and the dilution rate effect matrix R have similar structures.

[0051] Cross-interference correction in S70:

[0052] 1. Construct the cross-interference feature matrix X:

[0053]

[0054] In the formula, x ij This represents the interference coefficient of the i-th cytokine on the j-th cytokine.

[0055] 2. Cross-interference correction calculation:

[0056]

[0057] In the formula, S corrected The corrected signal matrix; S raw is the original signal matrix; I is the identity matrix; ° represents the Hadamard product; β is the interference intensity coefficient; γ is the correction bias term.

[0058] Data standardization in S80:

[0059] 1. Construct the original detection signal into a three-dimensional sparse matrix M. original :

[0060] M original ={m ijk} l×n×t ;

[0061] In the formula, m ijk Let l represent the signal value of the j-th cytokine in the i-th sample at the k-th time point, where l is the number of samples, n is the number of cytokine types, and t is the number of detection time points.

[0062] 2. Block processing matrix representation:

[0063]

[0064] In the formula, Let h be the element value in the p-th block, h be the block height, and k be the block width.

[0065] 3. Mathematical expression of the feature fusion algorithm:

[0066]

[0067] In the formula, F i G is the signal feature vector. j H is the background feature vector. k Let α be the noise feature vector. i ,β j ,γ k ∈ represents the corresponding weight coefficient, and ∈ represents the fusion error term.

[0068] 4. Correction signal calculation:

[0069] S final =M block ×F fusion ×(I+K compensation )×e -λt ;

[0070] In the formula, K compensationLet λ be the compensation coefficient matrix, λ be the time decay coefficient, and t be the detection time.

[0071] The evaluation system in S90:

[0072] 1. Construct a quality control evaluation matrix Q:

[0073]

[0074] In the formula, q ij This represents the evaluation index of the i-th quality control product for the j-th cytokine.

[0075] 2. Precision evaluation calculation:

[0076]

[0077] In the formula, CV ij σ is the coefficient of variation. ij The standard deviation is μ. ij δ is the average value. ij This represents systematic error.

[0078] 3. Accuracy evaluation equation:

[0079]

[0080] In the formula, A ij For accuracy, C measured To measure concentration, C theoretical For the theoretical concentration, η T ,η H ,η R These are error terms introduced by temperature, humidity, and dilution rate, respectively.

[0081] 4. Calculation of comprehensive evaluation indicators:

[0082]

[0083] In the formula, w ij For evaluation weights, θ is the weight coefficient of the determinant of the quality control matrix, and φ is the evaluation correction term.

[0084] Explanation of the principle behind the above equations:

[0085] 1) The three-dimensional sparse matrix representation fully considers the correlation between the three dimensions of sample, cytokine type and time;

[0086] 2) Block matrix processing can effectively reduce computational complexity and improve computational efficiency;

[0087] 3) Feature fusion adopts a weighted summation form, while introducing nonlinear terms to capture complex relationships;

[0088] 4) Time decay term e-λt This reflects the changing pattern of sample stability over time;

[0089] 5) The determinant term det(Q) is introduced into the evaluation system to characterize the integrity and stability of the quality control system.

[0090] The derivation process for each equation is explained below:

[0091] 1. Derivation of the multivariate mixed decomposition equation of the standard curve:

[0092] M raw =M stable +M var +M noise ;

[0093] Based on signal processing theory, signal decomposition is achieved through the following steps:

[0094] Step 1: Perform multi-scale analysis on the original signal using wavelet transform, and represent the signal as:

[0095]

[0096] Parameter explanation: W ψ f(a,b): Wavelet transform coefficients; a: Scaling coefficient, controlling the scaling of the wavelet, usually taken as an integer power of 2; b: Translation coefficient, controlling the translation position of the wavelet; ψ: Wavelet basis function, usually Daubechies wavelet; f(t): Input original signal; t: Time variable;

[0097] Step 2: Construct an EMD decomposition model and extract the intrinsic mode functions of the signal:

[0098]

[0099] Parameter explanation: e upper (t): Upper envelope function; e lower (t): Lower envelope function; m1(t): Mean envelope; a i ,b i ,c i ,d i : The cubic spline interpolation coefficients of the upper envelope are determined by boundary conditions; p i ,q i ,r i ,s i : Lower envelope cubic spline interpolation coefficients; n: Number of interpolation intervals, determined by the number of extreme points;

[0100] Step 3: Establish the objective function for reconstruction and optimization:

[0101]

[0102] Parameter explanation: ||·|| F Frobenius norm: used to measure the overall difference between matrices; ||·|| * λ1: Kernel norm, used to extract low-rank structural features; ||·||1: L1 norm, used to promote sparsity; ||·||2: L2 norm, used to control noise amplitude; λ1, λ2, λ3: Weighting coefficients, determined through cross-validation, typically ranging from 0.01 to 1; M raw : Original signal matrix; M atable : Stable signal component matrix; M var : Variable signal component matrix; M noise : Noise component matrix;

[0103] 2. The process of establishing the calibration curve equation:

[0104]

[0105] Step 1: Establish the relationship between fluorescence intensity and concentration based on Lambert-Beer's law:

[0106] F = F0(1-e -εlC );

[0107] Parameter explanation: F: fluorescence intensity; F0: incident light intensity, obtained through instrument calibration; ε: molar absorptivity, a material property parameter, in L / (mol·cm); l: optical path length, i.e., the thickness of the sample cell, in cm; C: analyte concentration, in mol / L;

[0108] Step 2: Perform Taylor expansion in the low concentration range and introduce the instrument response function:

[0109] s ij =αF + βF 2 +γF 3 ;

[0110] Parameter explanation: s ij : Stable signal response value; α: First-order response coefficient, obtained through standard calibration; β: Second-order response coefficient, reflecting nonlinear effects; γ: Third-order response coefficient, reflecting higher-order nonlinear effects;

[0111] Step 3: Establish the characteristics of the fluctuating signal through power spectrum analysis.

[0112]

[0113] Parameter explanation: S vv (f): Power spectral density function; Variance of the variable signal; τ0: characteristic correlation time, typically on the order of milliseconds; f: frequency variable; f c Cutoff frequency, determined by system bandwidth; v ij : Variable signal components;

[0114] 3. Derivation of the cross-interference correction equation:

[0115]

[0116] Step 1: Establish a cross-interference network model:

[0117]

[0118] Parameter explanation: k1, k2 are kinetic rate constants (determined through kinetic experiments), K d The dissociation constant (reflecting the bonding strength), K m Michaelis constant (characterizing reaction properties), c i ,c j For the concentrations of the i and j cytokines, f ij This is the interaction strength function.

[0119] Step 2: Constructing a competitive inhibition model:

[0120]

[0121] Parameter explanation: α i Intrinsic growth rate (determined by cytokine properties), β ij The competition coefficient (reflecting the intensity of inhibition), μ i Natural decay rate (related to molecular stability), S i denoted as the signal intensity of the i-th cytokine.

[0122] Step 3: Optimize the objective function design:

[0123]

[0124] Parameter explanation: and λ1 and λ2 are the corrected and true signal values, respectively; λ1 and λ2 are regularization parameters (determined through cross-validation); β is the interference intensity coefficient (range 0.1 to 0.5); and λ is the bias correction term.

[0125] 4. Derivation of the feature fusion algorithm:

[0126] Step 1: Constructing multidimensional feature vectors:

[0127]

[0128] Parameter explanation: s(t) is the signal time series, b(t) is the background time series, T is the total sampling time, t1, t2 are the integration intervals, FFT is the Fast Fourier Transform, σ 2 This is the variance operator.

[0129] Step 2: Design an adaptive weight update mechanism:

[0130]

[0131] η(t)=η0(1+μt) -v ;

[0132] Parameter explanation: To integrate features, For true features, λ α η is the regularization coefficient, η0 is the initial learning rate (usually 0.01), and μ and v are the decay parameters (0.1 and 0.5 respectively).

[0133] 5. Error Analysis and Uncertainty Assessment:

[0134] Calculation of combined standard uncertainty:

[0135]

[0136] Parameter explanation: u A Type A standard uncertainty (obtained from repeated measurements), u B Type B standard uncertainty (determined by instrument accuracy) For sensitivity coefficient, u(x) i ,x j ) represents the covariance between input quantities.

[0137] Final expanded uncertainty: U = ku c ;

[0138] Parameter explanation: k is the coverage factor (usually 2, corresponding to a 95% confidence level), and U is the expanded uncertainty (the measurement uncertainty in the final report).

[0139] Compared with existing technologies, the beneficial effects of the multiplex quantitative detection method for cytokines provided by this invention are:

[0140] 1. Standardization of the testing process has been achieved. This method adopts a systematic standard operating procedure in key steps such as sample pretreatment, standard preparation, and detection parameter optimization, ensuring the comparability and repeatability of test results across batches and significantly improving the reliability of the test results.

[0141] 2. Cross-interference between cytokines is eliminated. This method constructs a mathematical model of the interactions between cytokines and effectively compensates for the systematic bias caused by cross-reactions through a dynamic correction algorithm, thereby improving the accuracy of detection.

[0142] 3. Intelligent multi-factor detection is achieved. This method employs block matrix processing and feature fusion algorithms, enabling efficient and accurate simultaneous detection of multiple cytokines, significantly improving detection time efficiency and throughput.

[0143] 4. A comprehensive quality control system was established. This method designed a three-level quality control mechanism, including local feature analysis, global performance evaluation, and comprehensive index monitoring, to ensure the stability of the testing process and the reliability of the test results.

[0144] Therefore, this invention solves the problem of low accuracy in existing multiplex quantification techniques for cytokines. Attached Figure Description

[0145] Figure 1 A flowchart of the method provided by the present invention;

[0146] Figure 2 This is a decomposition diagram of the cytokine detection signal in Example 2;

[0147] Figure 3 This is a quality control evaluation feature analysis diagram from Example 2. Detailed Implementation

[0148] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings.

[0149] like Figure 1 The diagram shown is a flowchart of a multiplex quantification method for cytokines provided by this invention. The specific implementation details of each step are described below:

[0150] Step S10: Standardized preprocessing of cytokine detection samples. The purpose of this step is to perform standardized preprocessing of the samples, ensuring the standardization of sample separation and dilution processes. First, depending on the sample type, centrifugation is used to separate the samples at 500-5000 rpm for 5-30 minutes. Then, the supernatant is diluted at different ratios from 1:2 to 1:100, with three parallel samples for each dilution. Next, a standardized acquisition system is used to record various signal data in real time during the detection process, including the fluorescence signal intensity of the target cytokine, background signal intensity, and internal control signal intensity. Simultaneously, the operating parameters of the detection system, such as detection temperature, sample dilution rate, and system response time, are also recorded. Finally, a data preprocessing workflow is established to perform baseline drift correction, signal smoothing, and peak identification on the acquired signal data. These preprocessing steps ensure the reliability and consistency of the data in subsequent detection processes.

[0151] Step S20: Preparation of Cytokine Standard Sequences. The core of this step is establishing a concentration gradient sequence of standards, which serves as the basis for subsequent calibration of detection results. First, the lyophilized cytokine standard stock solution is dissolved in ultrapure water, and the concentration is converted according to the molecular weight. Then, cytokine standards are prepared using a standard lyophilization process. By controlling parameters such as the temperature gradient, vacuum degree, and amount of preservative added during the lyophilization process, the quality and stability of the standards are ensured. Next, the self-made standards are calibrated using international standards, the relative calibration factor is calculated, and the conversion relationship between standard concentration and international units is established. Finally, an 8-point concentration gradient standard dilution sequence is prepared, with adjacent concentration points diluted by a factor of 2, to cover the linear detection range of the detection system. The standard sequence data is converted into a sparse matrix form to prepare for the subsequent calibration model construction.

[0152] Step S30: Constructing a calibration model for cytokine detection. The core of this step is to extract stable and variable components of the signal based on standard data using a multivariate hybrid decomposition method, thereby establishing a calibration curve model. First, the standard data is constructed as a sparse matrix, with rows representing the standard concentration gradient and columns representing the cytokine types. Next, multi-scale decomposition methods such as wavelet transform and empirical mode decomposition are used to adaptively decompose this sparse matrix, extracting stable and variable signal components of different frequencies. Then, independent component analysis is used to further optimize the decomposition results, establishing a mathematical model for the signal components. Finally, a reconstruction optimization objective function is constructed and solved using an iterative algorithm to obtain the final decomposition result. In addition to the decomposition of the standard signal, this step also includes performing the same multivariate hybrid decomposition on the background signal during the detection process, laying the foundation for subsequent background signal correction.

[0153] Step S40: Optimize the conditions for multiplex cytokine detection. The key to this step is to use orthogonal experimental design to determine the optimal combination of operating parameters for the reaction system. First, an orthogonal experimental scheme is designed, selecting a reaction temperature range of 20-37 degrees Celsius, an incubation time range of 1-3 hours, and an oscillation frequency range of 200-800 rpm. Then, analysis of variance is used to calculate the contribution of each factor to the detection results, determining the main effect and interaction effect. Next, a factor weight calculation model based on the entropy weight method is established to determine the optimal weight coefficients for each parameter. Then, a parameter optimization objective function is constructed, and the optimal parameter combination is solved using nonlinear programming. Finally, the stability and reproducibility of the optimized parameter combination are evaluated through experiments. These steps ensure the stability and repeatability of the detection process.

[0154] Step S50: Establish an inter-batch variance correction model. The purpose of this step is to eliminate potential systematic biases between different batches. First, a sparse matrix of batch variances is constructed, and the within-batch and between-batch variances are calculated. Then, principal component analysis is used to perform multivariate decomposition of the variance matrix to extract the main sources of variation. Next, a weight determination model is established based on the entropy method to determine the influence weights of each source of variation. Optionally, Fisher's information matrix is ​​then introduced to evaluate the sensitivity of the parameters and optimize the batch correction coefficient matrix. Finally, cross-validation is used to evaluate the stability and generalization of the correction model. These steps ensure the comparability of test results between different batches.

[0155] Step S60: Design a three-tiered quality control system for multiplex cytokine detection. The core of this step is establishing a complete quality control monitoring mechanism to ensure the stability of the detection process. First, quality control samples at three concentration levels (high, medium, and low) are prepared, covering the typical concentration range of the target cytokines. Then, a local feature extraction model is established, including signal peak shape features, baseline features, and noise features, to evaluate the local stability of the detection system. Next, a global feature analysis model is constructed, including system stability, linear range, and reproducibility indicators, to evaluate the overall performance of the detection process. Then, a weighted summation feature fusion algorithm is used to comprehensively analyze local and global features. Finally, a quality control early warning mechanism is established, setting early warning thresholds and alarm rules to achieve full monitoring of the detection process. These steps ensure the reliability and credibility of the detection results.

[0156] Step S70: Establish a model to eliminate cross-interference between cytokines. The purpose of this step is to compensate for the mutual interference effects between different cytokines. First, a sparse matrix of cross-interference features is constructed to quantitatively describe the interaction strength between various cytokines. Then, a competitive inhibition model is used to describe the interaction relationships between cytokines, and the steady-state solution is calculated to obtain the cross-interference matrix. Next, an alternating minimization algorithm is used to optimize the correction parameters to minimize the impact of cross-interference. Finally, an iterative optimization program is established to dynamically adjust the cross-interference correction coefficients, ensuring the stability of the correction results. These steps effectively eliminate the mutual interference between different cytokines and improve the accuracy of the detection results.

[0157] Step S80: Standardize the detection data. The core of this step is to employ a block matrix processing strategy combined with a multi-dimensional signal feature fusion algorithm to generate the final cytokine concentration quantification result. First, the original detection signal is constructed into a three-dimensional sparse matrix, fully considering the correlation between sample, cytokine type, and time dimensions. Then, a block processing strategy is used to efficiently compute large-scale data, improving computational speed. Next, a feature fusion algorithm based on weighted summation is designed to integrate multi-dimensional detection signal features, including signal features, background features, and noise features. Afterward, a calibration parameter optimization model is established to achieve adaptive calibration of the detection signal. Finally, the final cytokine concentration quantification result is generated, and the accuracy of the result is evaluated. These steps ensure the standardization of the detection data and improve the reliability of the final result.

[0158] Step S90: Establish an evaluation system for cytokine detection results. The purpose of this step is to establish a complete quality evaluation mechanism to comprehensively assess the precision and accuracy of the detection results. First, a quality control evaluation matrix is ​​constructed, defining the evaluation indicators corresponding to each quality control product. Then, the weights of each evaluation indicator are determined using the analytic hierarchy process (AHP), and the weight coefficients are optimized using simulated annealing to achieve dynamic adjustment of the indicator weights. Next, the coefficient of variation and relative error of the detection results are calculated to evaluate precision and accuracy respectively. Finally, by integrating the characteristics of the quality control matrix, a comprehensive evaluation index is established to achieve a comprehensive quality evaluation of the detection results. These steps ensure the reliability and traceability of the detection results, providing support for subsequent clinical applications.

[0159] Specifically, the principle of this invention is:

[0160] 1. Standardized Pretreatment: A standardized pretreatment process involving centrifugation and dilution is employed to ensure consistency in sample processing and provide reliable baseline data for subsequent testing. Simultaneously, various operational parameters during the testing process are recorded in real time, providing a basis for subsequent calibration and compensation.

[0161] 2. Preparation of standards: Cytokine standards were prepared using a freeze-drying process and calibrated using international standards to establish a quantitative relationship between standard concentration and detection signal, laying the foundation for the construction of standard curves.

[0162] 3. Calibration Model Construction: A multivariate hybrid decomposition method is employed to decompose the standard sample detection signal into stable, variable, and noise components. Based on the decomposition results, a calibration curve model is established, encompassing factors such as signal stability, linear range, and sensitivity, which can accurately describe the relationship between standard concentration and detection signal.

[0163] 4. Optimization of detection conditions: Orthogonal experimental design was applied to optimize parameters such as temperature, time and oscillation frequency of the reaction system to determine the combination of process parameters that can maximize detection sensitivity and reproducibility.

[0164] 5. Batch variation correction: A batch variation matrix was constructed, and multivariate decomposition was used to extract the main sources of variation, effectively eliminating systematic bias between different batches.

[0165] 6. Cross-interference elimination: A mathematical model of the interaction between cytokines is established, and the cross-interference coefficient matrix is ​​calculated through a dynamic correction algorithm to compensate and correct the original detection signal in order to eliminate the error caused by cross-interference.

[0166] 7. Data Standardization: The original detection signal is constructed into a three-dimensional sparse matrix. A block matrix processing strategy and feature fusion algorithm are adopted to achieve efficient processing and adaptive correction of multi-dimensional detection data, and generate the final cytokine quantification results.

[0167] 8. Quality Control: A three-level quality control system is designed, including local feature analysis, global performance evaluation, and comprehensive index monitoring. The feature fusion algorithm enables real-time monitoring and quality evaluation of the testing process, ensuring the reliability of the test results.

[0168] The following is a specific embodiment 1 of the present invention. The specific implementation details of the steps in this embodiment 1 are described below: Step S10: First, determine the centrifugation parameters according to the sample type. For samples, use v = 3000 rpm and t = 15 min; for tissue homogenate samples, use v = 4500 rpm and t = 25 min. Sample dilution is performed using a serial dilution method, with the dilution factor satisfying D... i =2 i ,i∈[1,7]. When the acquisition system records the signal, the fluorescence signal is fitted using F(t)=F0+Asin(ωt+φ)+ΔF(t), where A is the amplitude, ω is the frequency, and φ is the phase; the background signal is fitted using B(t)=B0+k1t+k2t. 2Fitting was performed; the intrinsic parameter signal was corrected using I(t) = I0(1 + αT + βH + γR). Temperature control accuracy ΔT = ±0.1℃, humidity control accuracy ΔH = ±1%. Wavelet transform was used for signal preprocessing. Noise reduction is performed.

[0169] Specific implementation method of step S20: When dissolving the standard, the following method is used: Calculate the concentration. Temperature control during freeze-drying process: T(t) = T0e -kt Initial temperature T0 = -40℃, k = 0.1h -1 Vacuum degree control P(t) = P0(1-e -λt The target vacuum level is P0 = 5 Pa. Calibration relative to international standards is performed using... The concentration gradient of the 8-point standard curve is i∈[0,7]pg / mL. When constructing the standard data matrix, each concentration point was measured three times, and the average value was taken.

[0170] Step S30 Specific Implementation: Standard Product Data Matrix M raw Before performing multivariate mixture decomposition, normalization is performed first. Stable component extraction is adopted in The Laplace operator is used. The variational components use M. var =m raw -M stable -λΔM is extracted, where ΔM is the time difference. The noise component is filtered through a high-frequency filter M. noise =M raw *G σ Obtain, G σ The kernel function is Gaussian. The calibration curve fitting uses weighted least squares, with weights... Where σ i The standard deviation is denoted as .

[0171] Step S40 Specific Implementation: Orthogonal experimental design using L9(3 4 An orthogonal array is used, with temperature level T∈{25,30,35}℃, time level t∈{1,2,3}h, and oscillation frequency f∈{400,500,600}rpm. Contribution analysis is performed using... Among them SS i S is the sum of squares of the i-th factor. T This represents the total sum of squares. The weighting coefficients are calculated using the entropy weighting method. Where H i Let be the entropy value of the i-th index. The optimal parameter combination is solved using the response surface methodology.

[0172] Detailed implementation of step S50: When constructing the batch difference matrix, a reference batch normalization method is used. When principal component analysis extracts the main sources of variation, the eigenvalue decomposition ∑=UΛU T Principal components with a cumulative contribution rate greater than 85% were selected. Weights were determined using the entropy method, and eigenvectors were used. Fisher Information Matrix Used to evaluate parameter sensitivity. Cross-validation employs the k-fold method, where k = 5.

[0173] Step S60 Specific Implementation Method: The quality control product is prepared using C H =0.8C max C M =0.5C max C L =0.2C max Local feature extraction includes peak shape feature F. peak ={h,w,A,S}, baseline feature F base ={drift,noise,trend}. Global feature analysis uses principal component scores. The feature fusion algorithm uses weighted summation. The weights are determined using the AHP method. The warning thresholds are set according to the 3σ principle: UCL = μ + 3σ, LCL = μ - 3σ.

[0174] Step S70 Specific Implementation: The cross-interference matrix is ​​suppressed through a competitive suppression model. Construction. The steady-state solution is obtained by solving for x using Newton's iteration method. k+1 =x k -[J(x k )] -1 f(x k The correction parameters are optimized using the Alternating Direction Multiplier Method (ADMM): x k+1 =argmin x L ρ (x,z k ,y k ), z k+1 =argmin z L ρ (x k+1 ,z,y k ), y k+1 =y k +ρ(Ax k+1 +Bz k+1 -c).

[0175] Step S80 Specific Implementation: Three-dimensional sparse matrix block processing adopts Feature fusion employs a deep learning model F fusion =σ(W2σ(W1F+b1)+b2), where σ is the activation function. The calibration parameters are optimized using the Adam algorithm: m t =β1m t-1 +(1-β1)g t ,

[0176] Step S90 Specific Implementation: Quality control evaluation indicators include precision Accuracy linear correlation coefficient Weight optimization uses the simulated annealing algorithm: Temperature decay T k =αT k-1 The final evaluation index adopts a comprehensive scoring method. Where X i These are the standardized evaluation indicators.

[0177] To better understand and implement this invention, the following is a specific application scenario of this invention, Example 2:

[0178] A researcher is conducting a clinical study on the inflammatory response to immune infectious diseases, which requires the detection and analysis of multiple cytokines in the serum. To improve the accuracy and efficiency of the detection, the researcher decided to use the standardized method for multiplex cytokine detection proposed in this invention.

[0179] The first step involved standardized pretreatment of the serum samples to be tested. The collected serum samples were placed in centrifuge tubes and centrifuged at 3000 rpm for 10 minutes. The resulting supernatant was diluted 1:2, 1:5, and 1:10, with three replicates for each dilution. Simultaneously with signal acquisition, the detection temperature (room temperature 25°C) and sample dilution rates were recorded in real time. To eliminate baseline drift and noise interference, the researchers performed baseline correction and smoothing on the acquired fluorescence signal data.

[0180] In the second step, the researchers prepared cytokine standards. They first prepared stock solutions of 1 representative cytokines (IL-2, IL-4, IL-6, TNF-α, IFN-γ, IL-5, IL-1β, IL-8, IL-10, IFN-α, IL-17, and IL-12p70) using ultrapure water and converted the concentrations according to molecular weight. Then, they prepared these cytokine standards using a standard freeze-drying process, controlling the temperature gradient within 1℃ / min and the vacuum degree below 10 Pa, and adding 5% sucrose as a preservative. Finally, they calibrated using international standards, calculated the relative calibration factor, and established the conversion relationship between standard concentrations and international units. They prepared standard dilution sequences with eight concentration gradients (0.1, 0.5, 1, 5, 10, 20, 50, and 100 pg / mL), with three parallel samples at each concentration point. These standard data were organized into a sparse matrix to prepare for subsequent calibration model construction.

[0181] The third step involved researchers constructing a calibration model for cytokine detection. First, they used a matrix of standard sample data, M... raw Perform multivariate mixture decomposition to extract the stable component M of the signal. stable Variable component M var and noise component M noise :

[0182]

[0183] Where, f ij This represents the original fluorescence signal value of the j-th cytokine at the i-th concentration point.

[0184] Next, based on the decomposition results, the researchers established the calibration curve equation:

[0185]

[0186] Among them, C ij Let s be the standard concentration of the j-th cytokine at the i-th concentration point. ij ,v ij ,n ij Let k1, k2, and k3 be the stable component, the variable component, and the noise component, respectively; k1, k2, and k3 be the fitting coefficients; b be the intercept; and ∈ be the fitting error.

[0187] Meanwhile, the researchers also performed a similar multivariate decomposition on the background signal, obtaining the stable component M of the background signal. background,stable and variation component M background,var .

[0188] like Figure 2As shown in the figure, the original detection signal and its decomposed components are illustrated, including stable, variable, and noise components. The horizontal axis represents cytokine concentration (pg / mL), and the vertical axis represents fluorescence signal intensity. Different curves illustrate the various components of the signal, aiding in understanding the signal decomposition process in step three.

[0189] Fourthly, to optimize the reaction conditions for cytokine detection, the researchers designed an orthogonal experimental design. They selected reaction temperature (20-37℃), incubation time (1-3h), and oscillation frequency (200-800rpm) as optimization factors. Through analysis of variance, they calculated the main effects and interaction effects of each factor, finding that temperature and incubation time were the two main factors affecting the detection results.

[0190] Then, based on the entropy weight method, the researchers constructed the following factor weight calculation model:

[0191]

[0192] Where, w i p is the weight coefficient of the i-th factor. ij This represents the result of the j-th level of the i-th factor after normalization.

[0193] Finally, they established the following objective function for parameter optimization and used nonlinear programming to solve for the optimal parameter combination:

[0194] x * =argmin x F obj (x);

[0195]

[0196] Where, x * For the optimized parameter combination, The expected detection results were obtained. Through experimental verification, the researchers determined the final reaction conditions to be: temperature 30℃, incubation time 1h, and oscillation frequency 400rpm.

[0197] Fifth, to eliminate potential systematic biases between different batches, the researchers established a batch-to-batch difference correction model. First, they constructed a batch-to-batch difference matrix D and calculated the differences in the detection of 12 cytokines among the five batches, as shown in Table 1:

[0198] Table 1 Batch Differences

[0199]

[0200] IL-2 0.12 0.08 0.05 0.15 0.09 IL-4 0.14 0.11 0.07 0.13 0.10 IL-6 0.16 0.09 0.06 0.12 0.11 TNF-α 0.18 0.13 0.08 0.14 0.12 IFN-γ 0.15 0.10 0.09 0.16 0.13 IL-5 0.13 0.12 0.08 0.14 0.11 IL-1β 0.17 0.09 0.07 0.13 0.10 IL-8 0.14 0.11 0.06 0.15 0.12 IL-10 0.16 0.10 0.08 0.12 0.09 IFN-α 0.15 0.12 0.07 0.14 0.11 IL-17 0.13 0.09 0.06 0.16 0.10 IL-12p70 0.17 0.11 0.08 0.13 0.12

[0201] Then, the researchers calculated the batch correction coefficient matrix K:

[0202] K = D·W·T·H·R;

[0203] Where W is the weight matrix, and T, H, and R are the influence matrices of temperature, humidity, and dilution rate, respectively. Through cross-validation, the researchers evaluated the stability and generalization ability of the calibration model.

[0204] Step six: To eliminate cross-interference among cytokines, the researchers constructed a cross-interference correction model. First, they established a cross-interference feature matrix X, which quantitatively described the interaction strength among the 12 cytokines, as shown in Table 2:

[0205] Table 2 Cross-interference characteristics:

[0206]

[0207] Then, the researchers used a competitive inhibition model to describe the interactions between cytokines and calculated the steady-state solution X. * :

[0208]

[0209]

[0210] Where, S i Let k be the concentration of the i-th cytokine. i and k ij These are the self-elimination rate and the mutual inhibition rate, respectively.

[0211] Finally, the researchers used an alternating minimization algorithm to optimize the correction parameters α, β, and γ, and compensated for and corrected the original detection signal.

[0212]

[0213] In the seventh step, to standardize the detection data, the researchers first constructed the original signal into a three-dimensional sparse matrix M. original Its dimensions are 100×5×6, representing the signal values ​​of 100 samples, 12 cytokines, and 6 detection time points.

[0214] Then, a block processing strategy is adopted to divide the matrix into multiple sub-block matrices. As shown in Table 3:

[0215] Table 3 Samples of block processing

[0216]

[0217] Next, the researchers designed a feature fusion algorithm based on weighted summation to integrate multi-dimensional detection signal features:

[0218]

[0219] Where, F target,i For the signal characteristics of the i-th cytokine, F background,j For the j-th background feature, F noise,k For the k-th noise feature, α i ,β j ,γ k ∈ represents the corresponding weight coefficient, and ∈ represents the fusion error.

[0220] Finally, the researchers calculated the final correction signal S. final :

[0221] S final =M block ·F fusion ·(I+K compensation )·e -0.05t ;

[0222] Where, K compensation Here is the compensation coefficient matrix, and λ = 0.05 is the time decay coefficient.

[0223] Step 8: To comprehensively evaluate the quality of the test results, the researchers established the following quality control evaluation system. First, they prepared quality control samples Q at three concentration levels (1000, 200, and 100 pg / ml), covering the typical concentration range of clinical samples, as shown in Table 4:

[0224] Table 4 Concentration of Quality Control Products

[0225]

[0226]

[0227] Then, the researchers established a local feature extraction model, including the signal peak shape feature F. peak Baseline characteristics F baseline and noise characteristics F noise :

[0228] F peak =max(F target )-min(F target );

[0229]

[0230] At the same time, they also constructed a global feature analysis model, including system stability F. stability Linear range Flinearity and reproducibility F reproducibility :

[0231]

[0232] Where, σ ij and μ ij , respectively, are the standard deviation and mean of the j-th cytokine in the i-th batch.

[0233] Finally, the researchers used a weighted summation method to integrate local and global features and established a comprehensive evaluation index E. total :

[0234]

[0235] Where, w ij These are the weighting coefficients for each feature.

[0236] like Figure 3 The image shows the scores of different cytokines on various quality control characteristics. The evaluation results include six dimensions such as peak shape and baseline characteristics.

[0237] Through the standardized procedures outlined above, researchers obtained quantitative results for 12 cytokines in the serum. After quality control evaluation, the results demonstrated good precision and accuracy, meeting the requirements for clinical application. This provides crucial biomarker data support for subsequent research on immune-mediated infectious diseases.

[0238] It should be noted that the variables involved in the equations of this invention are explained in detail in Table 5 below.

[0239] Table 5. Explanation of Equation Variables

[0240]

[0241]

[0242] It should be noted that the variables involved in the derivation of this invention are explained in detail in Table 6 below.

[0243] Table 6. Explanation of Variables in the Derivation Process

[0244]

[0245] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for multiplex quantitative detection of cytokines, comprising the following steps: Standardized preprocessing was performed on cytokine detection samples. Fluorescence signal data, background signal data, standard concentration data, sample dilution rate, reaction temperature, and ambient humidity were collected in real time during the detection process. Standard sequences were prepared using a standard lyophilization process for cytokine standards. A calibration model was constructed, converting the standard data into a sparse matrix and performing multivariate mixture decomposition to obtain stable and variable signal components. The calibration coefficient matrix was calculated based on the decomposition results. Orthogonal experiments were used to optimize the multiplex detection conditions for cytokines. A batch-to-batch difference correction model was established, constructing a sparse matrix of batch-to-batch differences and performing multivariate mixture decomposition to obtain stable and variable components. A three-level quality control system was designed, preparing high-concentration, medium-concentration, and low-concentration quality control samples. A signal cross-interference elimination model was established, constructing a sparse matrix of cross-interference features and performing multivariate mixture decomposition. The detection data were standardized. A detection result evaluation system was established, and the final concentration of cytokines in the samples was calculated.

2. The method for multiplex quantitative detection of cytokines according to claim 1, characterized in that, Standardized preprocessing of cytokine detection samples includes: collecting samples for centrifugation, with centrifugation speeds ranging from 500 to 5000 rpm and centrifugation times from 5 to 30 minutes, depending on the sample type; serially diluting the supernatant after separation, with dilution ratios ranging from 1:2 to 1:100; using a standardized acquisition system to record fluorescence signal data in real time, including target signal intensity, background signal intensity, and internal reference signal intensity; synchronously acquiring detection system operating parameters; and establishing a data preprocessing workflow to perform baseline drift correction, signal smoothing, and peak identification on the acquired signals.

3. The method for multiplex quantitative detection of cytokines according to claim 2, characterized in that, The preparation of standard sequences includes: preparing stock solutions of cytokine standards; dissolving and lyophilizing the standards in ultrapure water; converting the concentrations according to molecular weight; preparing cytokine standards using a standard lyophilization process, controlling the temperature gradient, vacuum level, and amount of protective agent added during the lyophilization process; calibrating using international standards, calculating the relative calibration factor, and establishing the conversion relationship between standard concentrations and international units; preparing a standard dilution sequence with an 8-point concentration gradient, setting the dilution factor between adjacent concentration points to 2 times to cover the linear range of the detection system; and converting the standard sequence data into a sparse matrix form.

4. The method for multiplex quantitative detection of cytokines according to claim 3, characterized in that, The calibration model is constructed as follows: The standard data is constructed into a sparse matrix, where the number of rows in the sparse matrix represents the number of standard concentration gradients, and the number of columns represents the number of cytokine types; the sparse matrix is ​​decomposed into multiple scales, and wavelet transform is used to extract different frequency components, selecting the optimal number of decomposition levels; the signal is adaptively decomposed using empirical mode decomposition, extracting intrinsic mode functions and performing reconstruction optimization; independent component analysis is used to separate stable and variable components of the signal, establishing a mathematical model of the signal components; a reconstruction optimization objective function is constructed, and an iterative algorithm is used to optimize the reconstruction parameters.

5. The method for multiplex quantitative detection of cytokines according to claim 4, characterized in that, Orthogonal experiments were used to optimize the conditions for multiplex detection of cytokines, including: designing an orthogonal experimental scheme, selecting a reaction temperature range of 20 to 40 degrees Celsius, an incubation time range of 1 to 24 hours, and an oscillation frequency range of 200 to 800 rpm; calculating the contribution of each factor to the detection results, and using analysis of variance to determine the main effect and interaction effect; establishing a factor weight calculation model based on the entropy weight method to determine the weight coefficients of each parameter; constructing a parameter optimization objective function, and using nonlinear programming to solve for the optimal parameter combination; and verifying the stability and reproducibility of the optimized parameter combination.

6. The method for multiplex quantitative detection of cytokines according to claim 5, characterized in that, A batch-to-batch variance correction model was established, including: constructing a sparse matrix of batch-to-batch variances and calculating within-batch and between-batch variances; performing multivariate decomposition on the variance matrix and using principal component analysis to extract the main sources of variation; calculating the variance decomposition ratio and establishing a weight determination model based on the entropy method; introducing the Fisher information matrix to evaluate parameter sensitivity and optimize the correction coefficients; and using cross-validation to evaluate the stability of the correction model.

7. The method for multiplex quantitative detection of cytokines according to claim 6, characterized in that, The design of a three-level quality control system includes: preparing quality control samples at three concentration levels (high, medium, and low) and selecting typical concentration ranges for the target analyte; establishing a local feature extraction model, including signal peak shape features, baseline features, and noise features; constructing a global feature analysis model, including system stability, linear range, and reproducibility indicators; designing a feature fusion algorithm that integrates local and global features using a weighted summation method; and establishing a quality control early warning mechanism, setting early warning thresholds and alarm rules.

8. The method for multiplex quantitative detection of cytokines according to claim 7, characterized in that, A signal cross-interference cancellation model is established, including: constructing a sparse matrix of cross-interference features and establishing an interaction strength function; using a competitive inhibition model to describe the interaction relationship between cytokines; calculating the steady-state solution to obtain the cross-interference matrix and optimizing the nonlinear correction term; using an alternating minimization algorithm to optimize the correction parameters; and establishing an iterative optimization program to achieve dynamic parameter adjustment.

9. The method for multiplex quantitative detection of cytokines according to claim 8, characterized in that, The detection data is standardized, including: converting the original detection signal into a sparse matrix form and establishing a signal processing model; using a block processing strategy to perform efficient calculations on large-scale data; designing a feature fusion algorithm to integrate multi-dimensional detection signals; establishing a correction parameter optimization model to achieve adaptive signal correction; generating the final quantitative results and evaluating the accuracy of the results.

10. The method for multiplex quantitative detection of cytokines according to claim 9, characterized in that, Establish a testing result evaluation system, including: establishing a quality control product evaluation system and setting judgment standards for various indicators; using the analytic hierarchy process (AHP) to determine the weights of evaluation indicators; constructing a simulated annealing algorithm to optimize the weight coefficients; calculating the precision and accuracy of the system's testing results; and establishing a result reporting and quality traceability system.

Images