An immune turbidimetry-based method for detecting and analyzing nuclear matrix protein 22

Abstract

The application discloses a nucleomatrix protein 22 detection and analysis method based on an immune turbidimetry method, and relates to the technical field of immune detection. The method comprises the following steps: collecting an initial mixed solution of a to-be-detected sample and an antibody reagent, continuously monitoring a plurality of initial turbidity values within a preset reaction period; identifying an abnormal reaction period according to the initial turbidity values, and performing ultrasonic-assisted dispersion treatment on the initial mixed solution within the period to obtain a homogeneous mixed solution; adding an accelerated sedimentation medium, and collecting a plurality of dynamic turbidity values in a sedimentation process in real time; extracting a peak turbidity value and a corresponding peak time point, and calculating an initial concentration of the nucleomatrix protein 22; and inputting the initial concentration into a pre-trained nonlinear correction model to output a final concentration of the nucleomatrix protein 22. The method improves the accuracy and anti-interference capability of the immune turbidimetry method in detecting the nucleomatrix protein 22 in a complex sample through ultrasonic-assisted dispersion and nonlinear correction.

Description

Technical Field

[0001] This invention relates to the field of immunoassay technology, specifically to a method for detecting and analyzing nuclear matrix protein 22 based on immunoturbidimetry. Background Technology

[0002] Immunoturbidimetry is a technique for quantitatively analyzing analyte concentrations by measuring changes in turbidity in a solution after an antigen-antibody binding reaction. In the clinical detection of matrix protein 22, the main challenges of this method are the complexity of the sample matrix and the nonlinear nature of the immune reaction rate. Interfering substances in urine samples, such as cell debris, salt crystals, and mucins, can cause localized aggregation or uneven precipitation of immune complexes during formation, deviating the ideal homogeneous state of the reaction system. This manifests as irregular jumps or periodic fluctuations in the turbidity signal at the initial stage of the reaction. This signal abnormality disrupts the previously stable correlation between turbidity values ​​and antigen concentrations, distorting the underlying data for subsequent quantitative calculations. Existing solutions typically rely on extending incubation time, increasing mixing frequency, or performing complex sample pretreatments to attempt to allow the system to stabilize on its own. However, these passive measures increase overall detection time and fail to eliminate the persistent impact of existing localized aggregations on the overall optical signal. Another inherent challenge lies in the nonlinear relationship between immunoturbidity signals and antigen concentrations, especially when antigen concentrations are too high or too low. In these cases, the amount of immune complexes generated no longer maintains a linear correlation with concentration, a phenomenon known as the hook effect or pre- and post-zone phenomenon. Conventional detection methods rely on standard curve fitting, employing single linear or simple polynomial regression, which cannot accurately describe the concentration-turbidity response across the entire reportable range, leading to systematic biases in test results at extreme concentrations. Furthermore, this nonlinearity is amplified in the turbidity dynamics during the sedimentation phase. The timing and magnitude of peak turbidity are influenced by multiple factors, including sample matrix viscosity, antibody affinity, and temperature, making the process of deducing the initial concentration from peak parameters highly uncertain. The problem this invention aims to solve is eliminating signal disturbances in the initial reaction phase and correcting the nonlinear mapping from turbidity signals to concentration results. Summary of the Invention

[0003] The purpose of this invention is to provide a method for detecting and analyzing nuclear matrix protein 22 based on immunoturbidimetry, so as to eliminate signal disturbances caused by sample matrix interference in the early stage of the reaction, correct the nonlinear mapping deviation between turbidity signal and nuclear matrix protein 22 concentration, and improve detection accuracy.

[0004] The objective of this invention can be achieved through the following technical solutions:

[0005] This invention provides a method for detecting and analyzing nuclear matrix protein 22 based on immunoturbidimetry, comprising: collecting an initial mixture of a test sample and an antibody reagent, and continuously monitoring multiple initial turbidity values ​​of the initial mixture within a preset reaction period; identifying abnormal reaction periods of the initial mixture based on the multiple initial turbidity values, and performing ultrasonic-assisted dispersion treatment on the initial mixture during the abnormal reaction periods to obtain a homogeneous mixture; adding an accelerating sedimentation medium to the homogeneous mixture, and collecting multiple dynamic turbidity values ​​of the homogeneous mixture during the sedimentation process in real time; extracting the peak turbidity value and its corresponding peak time point from the multiple dynamic turbidity values, and calculating the initial concentration of nuclear matrix protein 22 based on the peak turbidity value and the peak time point; inputting the initial concentration into a pre-trained nonlinear correction model, and the nonlinear correction model outputting the final concentration of nuclear matrix protein 22. This method captures uneven mixing or abnormal aggregation by real-time monitoring of turbidity changes in the early stage of the reaction, and uses ultrasonic dispersion to quickly homogenize the reaction system, effectively eliminating turbidity fluctuations caused by local concentration differences and improving the stability of the detection signal. An accelerated sedimentation medium is introduced into the homogeneous mixture to promote rapid and orderly sedimentation of immune complexes, and the peak turbidity is accurately captured by continuously collecting dynamic turbidity data, avoiding measurement deviations caused by sedimentation time differences in traditional endpoint methods. Simultaneously, the initial concentration is corrected using a nonlinear correction model, which can compensate for nonlinear effects such as antigen excess and matrix interference, significantly improving the accuracy and anti-interference capability of the detection.

[0006] Preferably, the method for identifying the abnormal reaction period is as follows: Arrange the multiple initial turbidity values ​​into a turbidity time series in chronological order; calculate the difference between two adjacent initial turbidity values ​​in the turbidity time series to obtain multiple turbidity change rates; and filter out abnormal change rates whose absolute values ​​exceed a preset change rate threshold. The collection period corresponding to the abnormal change rate is marked as the abnormal reaction period. This allows for sensitive location of periods where turbidity exhibits non-physiological fluctuations, providing a clear time window for subsequent precise application of auxiliary dispersion treatment.

[0007] As another technical solution of the present invention, the specific process of performing ultrasound-assisted dispersion treatment during the abnormal reaction period is as follows: determining the maximum turbidity fluctuation amplitude in the initial mixture during the abnormal reaction period; selecting the initial operating frequency and initial operating power of the ultrasound transducer according to the maximum turbidity fluctuation amplitude; controlling the ultrasound transducer to emit ultrasound waves to the initial mixture according to the initial operating frequency and the initial operating power during the abnormal reaction period; and monitoring the real-time turbidity fluctuation amplitude of the initial mixture during the ultrasound treatment process in real time. When the real-time turbidity fluctuation amplitude drops below a preset fluctuation threshold, the ultrasound emission is stopped, resulting in a homogeneous mixture. Adaptively determining the ultrasound parameters based on the degree of turbidity fluctuation enables rapid dispersion of aggregates and elimination of concentration gradients without damaging antibody and antigen activity. Furthermore, terminating ultrasound through real-time feedback avoids overtreatment that could disturb subsequent immune responses. More preferably, the ultrasonic transducer simultaneously acquires ultrasonic echo signals while emitting ultrasonic waves, and adjusts the initial operating frequency and the initial operating power in real time according to the ultrasonic echo signals. By utilizing the echo characteristics to sense changes in the state of microbubbles or particles in the mixture, closed-loop control of ultrasonic energy is achieved, further ensuring the dispersion effect and the stability of the system temperature.

[0008] In a preferred embodiment of the present invention, the process of adding an accelerating sedimentation medium and collecting dynamic turbidity values ​​includes: selecting an accelerating sedimentation medium from a pre-stored sedimentation medium library that matches the current viscosity of the homogeneous mixture; the accelerating sedimentation medium being nanomagnetic beads with hydrophilic groups on their surface; adding the selected accelerating sedimentation medium to the homogeneous mixture at a preset dropping rate, while continuously stirring during the dropping process; and collecting multiple dynamic turbidity values ​​of the homogeneous mixture along the vertical optical path at fixed time intervals during the static sedimentation stage after the dropping is completed. By selecting nanomagnetic beads with hydrophilic surface modification and matching the viscosity of the mixture, non-specific adsorption can be avoided while enabling rapid and regular co-precipitation of the immune complex, thereby obtaining a turbidity decay curve with high signal-to-noise ratio and good repeatability. Preferably, the amount of accelerating sedimentation medium added is dynamically calculated based on the volume of the homogeneous mixture and the current viscosity, and the average particle size of the nanomagnetic beads is between 50 nm and 200 nm, balancing sedimentation speed and optical detection sensitivity, making the turbidity peak clearly distinguishable.

[0009] For calculating the initial concentration of nuclear matrix protein 22, this invention further proposes: plotting the multiple dynamic turbidity values ​​into a turbidity decay curve according to the chronological order of collection; identifying the inflection point where the turbidity value changes from rising to falling in the turbidity decay curve; determining the turbidity value at the inflection point as the peak turbidity value; and determining the time point corresponding to the inflection point as the peak time point; obtaining a pre-established reference mapping table of standard peak time points and standard concentrations; finding the standard peak time point closest to the peak time point in the reference mapping table; and reading the corresponding standard concentration as the initial concentration of nuclear matrix protein 22. This quantitative method based on the peak occurrence time avoids the problem of absolute turbidity being easily affected by optical and impurity interference, making the initial concentration calculation more robust.

[0010] To eliminate interference from nonlinear factors in detection, one technical solution of the present invention involves inputting the initial concentration into the nonlinear correction model as follows: acquiring multiple historical correction records stored in the nonlinear correction model, each historical correction record containing a historical initial concentration and a corresponding historical true concentration; calculating the concentration difference between the initial concentration and the historical initial concentration in each historical correction record; selecting matching correction records with concentration differences less than a preset difference threshold; weighting the historical true concentrations in the matching correction records; and using the weighted average result as the final concentration. This historical matching correction mechanism can perform real-time local correction of system deviations without complex iterations. As another preferred embodiment of the present invention, the training process for constructing the nonlinear correction model includes: collecting multiple nuclear matrix protein 22 standards with known true concentrations; obtaining the training initial concentration corresponding to each standard using the same calculation method as the initial concentration; forming training data pairs with the training initial concentration and the corresponding true concentration of each standard; using the training initial concentration as the input feature and the true concentration as the output label; training a feedforward neural network using the backpropagation algorithm until the loss function value converges to a preset loss threshold; and using the trained feedforward neural network as the nonlinear correction model. By learning the nonlinear mapping between the initial concentration and the true concentration through a data-driven approach, matrix effects and "hook effects" can be compensated more precisely, significantly improving detection accuracy in both high and low concentration ranges.

[0011] At the source of signal acquisition, this invention provides a preferred scheme for continuous monitoring of the initial mixture: The sample to be tested is injected into a reaction vessel containing antibody reagent according to a preset sample-to-reagent volume ratio; the reaction vessel is stirred at a first rotation speed for a first preset time under constant temperature conditions to obtain the initial mixture; within the preset reaction time period, a light source is controlled to vertically irradiate the initial mixture with a fixed light intensity, and a photodetector receives the transmitted light signal at a fixed sampling frequency; the transmitted light signal received at each sampling moment is converted into a corresponding turbidity value to obtain multiple initial turbidity values. The temperature-controlled mixing and standardized optical path design ensure the repeatability and comparability of the initial turbidity value acquisitions.

[0012] To improve sample applicability, a sample pretreatment step is included before collecting the initial mixture: A urine sample to be tested is obtained and placed in a centrifuge tube; the centrifuge is controlled to centrifuge the urine sample in the centrifuge tube at a second speed for a second preset time to remove precipitates and obtain supernatant; the total protein content of the supernatant is measured, and when the total protein content is below the protein threshold, the supernatant is used as the test sample. Centrifugation removes formed elements and excludes high-protein interfering samples, reducing the impact of non-specific turbidity on detection and ensuring the specificity of the test.

[0013] To ensure the reliability of the results, a retest can be performed after the final concentration is output: The pre-stored historical concentrations of the patient corresponding to the test sample are read, and the rate of change between the final concentration and the historical concentration is calculated. When the rate of change exceeds a preset threshold, a retest marker is generated, and the autosampler is controlled to re-extract a retest sample from the supernatant. The retest sample is then retested using the same calculation process as the final concentration to obtain the retest concentration, which replaces the final concentration. Automated retesting of abnormal fluctuations using individual historical data can reduce the risk of false positives or false negatives introduced by random errors.

[0014] The beneficial effects of this invention are:

[0015] By identifying anomalous periods in the initial mixture reaction process and performing ultrasonic-assisted dispersion treatment during these periods, the microjets and local shear forces generated by ultrasonic cavitation effectively break down and disperse local aggregates of immune complexes formed due to interference from the urinary matrix. The initial operating frequency and power of the ultrasonic transducer are dynamically selected based on the maximum turbidity fluctuation amplitude during these periods, ensuring that the applied acoustic energy matches the size and binding strength of the aggregates. This avoids poor dispersion due to insufficient energy or antibody inactivation due to excessive energy. Ultrasonic echo signals are simultaneously acquired and analyzed during ultrasonic emission to adjust the operating frequency and power parameters in real time, forming a closed-loop control that ensures the dispersion process always follows the real-time deagglomeration state of the aggregates. Ultrasonic treatment is immediately stopped when the real-time turbidity fluctuation amplitude drops below a preset fluctuation threshold, ensuring complete homogenization of the system and preventing damage to the formed antigen-antibody specific complexes from overtreatment. The resulting homogeneous mixture eliminates optical path obstruction and signal jumps caused by irregular precipitation, allowing subsequent turbidity acquisition during the sedimentation stage to be established on a stable reaction baseline. Nanomagnetic beads with hydrophilic surface-modified groups are added as an accelerating sedimentation medium. Their hydrophilic surfaces interact controllably with the immune complexes, forming a stable sedimentation interface under gravity. Optical signals are acquired vertically at fixed time intervals, and the resulting dynamic turbidity sequence accurately reflects the true sedimentation kinetics of the immune complexes in a homogeneous system, avoiding sedimentation curve distortion caused by heterogeneous interference. The initial concentration is calculated based on the peak turbidity value and peak time point, and this initial concentration is used as input by a nonlinear correction model pre-trained using historical correction records or a feedforward neural network. The model automatically compensates for the distortions in the turbidity-concentration relationship caused by hook effects and matrix effects by finding matching historical correction records and performing a weighted average of the historical true concentrations, or by learning and generalizing the nonlinear mapping relationship between training data pairs through a feedforward neural network. This accurately maps the initial concentration to the final concentration, overcoming the limitation of traditional linear fitting inaccuracies over extended concentration ranges. Attached Figure Description

[0016] The invention will now be further described with reference to the accompanying drawings.

[0017] Figure 1 This is a flowchart of the detection and analysis method for nuclear matrix protein 22 based on immunoturbidimetry;

[0018] Figure 2 This is a flowchart of the dynamic turbidity detection method for nuclear matrix protein 22 concentration;

[0019] Figure 3 This is a flowchart of the continuous monitoring process for initial turbidity values. Detailed Implementation

[0020] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] See Figure 1 This invention provides a method for detecting and analyzing nuclear matrix protein 22 based on immunoturbidimetry. The method includes: mixing the sample to be tested with an antibody reagent to form an initial mixture; continuously monitoring the initial mixture optically within a preset reaction period to obtain multiple initial turbidity values ​​distributed over time; identifying abnormal reaction periods where non-specific aggregation or abnormal fluctuations occur in the initial mixture based on the characteristics of the initial turbidity value changes; applying ultrasonic-assisted dispersion treatment to the initial mixture during these abnormal reaction periods using an ultrasonic transducer to eliminate large particle aggregation and restore the mixture to a homogeneous state; adding an accelerating sedimentation medium to the homogeneous mixture to accelerate the orderly sedimentation of the immune complexes; and acquiring multiple dynamic turbidity values ​​in real time along the vertical optical path during the sedimentation process; determining the peak turbidity value and its corresponding peak time point from the acquired dynamic turbidity values; and calculating the initial concentration of nuclear matrix protein 22 using the correlation between the peak time point and concentration; and inputting this initial concentration into a pre-constructed nonlinear correction model, which, combined with historical correction data, outputs the final concentration of nuclear matrix protein 22.

[0022] In practice, an initial mixture of the test sample and antibody reagent is collected, and multiple initial turbidity values ​​of the initial mixture are continuously monitored within a preset reaction period. This monitoring step can be completed using a light scattering turbidity detection system. The initial turbidity values ​​obtained from each sampling within the preset reaction period, along with the timestamp corresponding to the collection time, are stored in the memory of the data processing device. The data processing device arranges the multiple initial turbidity values ​​into a turbidity time series according to the chronological order of the timestamps. Each element in the turbidity time series contains a time point and the initial turbidity value corresponding to that time point.

[0023] In practical implementation, for a turbidity time series, the data processing device calculates the difference between each pair of adjacent initial turbidity values ​​and divides the difference by the time interval between two adjacent sampling times to obtain the turbidity change rate at the corresponding time. The turbidity change rate is calculated using the formula... To achieve, among which, Indicates the first The rate of change of turbidity calculated from each sampling interval Represents the turbidity time series of the first... An initial turbidity value, Represents the turbidity time series of the first... An initial turbidity value, This represents the time interval between two consecutive samples. The data processing device stores a preset rate of change threshold in a register and compares the absolute value of the calculated turbidity rate of change with the preset rate of change threshold one by one. If the absolute value of the turbidity rate of change exceeds the preset rate of change threshold, the turbidity rate of change is determined as an abnormal rate of change; if the absolute value of the turbidity rate of change does not exceed the preset rate of change threshold, the turbidity rate of change is determined as a normal rate of change. The preset rate of change threshold is a fixed value determined based on the kinetic characteristics of the immunoturbidimetric reaction and the noise level of the detection instrument. Its value is determined by: statistically analyzing the turbidity rate of change in the initial stage of multiple batches of normal immunoreactions, using the upper limit of turbidity fluctuation in normal reactions as a benchmark, and then combining it with the background noise amplitude of the photodetector to determine a safety margin, thereby setting the preset rate of change threshold to avoid misjudging noise as abnormal.

[0024] In practice, the data processing device marks the collection time period corresponding to the abnormal rate of change as the abnormal reaction period. The abnormal reaction period begins with the sampling interval where the first abnormal rate of change occurs and ends when no abnormal rate of change occurs in several consecutive sampling intervals. Within the determined abnormal reaction period, the data processing device reads all initial turbidity values ​​within that period, compares and searches for the maximum and minimum values ​​of the initial turbidity values, subtracts the maximum value from the minimum value, and determines the difference as the maximum turbidity fluctuation amplitude within the abnormal reaction period.

[0025] In practice, the data processing device selects the initial operating frequency and initial operating power of the ultrasonic transducer based on the maximum turbidity fluctuation amplitude. The selection method is as follows: an amplitude-parameter reference table is pre-established, dividing the table into several fluctuation amplitude intervals. Each fluctuation amplitude interval corresponds to a set of initial operating frequencies and initial operating powers. The larger the fluctuation amplitude interval value, the higher the corresponding initial operating frequency and initial operating power settings. The basis for establishing the amplitude-parameter reference table is that the cavitation shear force generated by the ultrasonic transducer's emitted sound field is sufficient to disperse aggregates of different sizes without damaging the antigen-antibody binding activity. A larger fluctuation amplitude indicates a larger aggregate size, requiring higher ultrasonic energy. By referring to the table, the data processing device selects the initial operating frequency and initial operating power corresponding to the fluctuation amplitude interval where the maximum turbidity fluctuation amplitude is located.

[0026] In practice, during the abnormal reaction period, the data processing device controls the ultrasonic transducer to emit ultrasonic waves into the reaction vessel containing the initial mixture at the selected initial operating frequency and power. The ultrasonic transducer is installed on the side wall of the reaction vessel, and the ultrasonic waves are transmitted into the initial mixture through a coupling medium. Simultaneously with the emission of ultrasonic waves, the data processing device continuously collects the turbidity value of the initial mixture using a turbidity detection system. Multiple turbidity values ​​are collected within a sliding time window, and the real-time turbidity fluctuation amplitude within that window is calculated. The real-time turbidity fluctuation amplitude is the difference between the maximum and minimum turbidity values ​​within that window. The data processing device compares the real-time turbidity fluctuation amplitude with a preset fluctuation threshold, which is set based on the inherent turbidity perturbation amplitude of a normal homogeneous mixture. When the real-time turbidity fluctuation amplitude drops below the fluctuation threshold, the data processing device immediately sends a stop signal to the ultrasonic transducer, which then stops emitting ultrasonic waves, resulting in a homogeneous mixture.

[0027] In practical implementation, the ultrasonic transducer, while emitting ultrasonic waves, also acts as a receiving transducer to simultaneously acquire ultrasonic echo signals reflected from scatterers in the mixture. The ultrasonic transducer converts the ultrasonic echo signals into electrical signals and transmits them to a data processing device. The data processing device extracts the intensity and frequency characteristics of the ultrasonic echo signals. The intensity characteristic is the moving average of the peak-to-peak voltage of the ultrasonic echo signal, and the frequency characteristic is the spectral centroid shift of the ultrasonic echo signal. Based on the difference between the intensity characteristics and a preset intensity target, and the difference between the frequency characteristics and a preset frequency target, the data processing device uses a proportional-integral adjustment algorithm to calculate the frequency correction and power correction in real time. The frequency correction is then added to the current operating frequency, and the power correction is added to the current operating power. This adjusts the operating parameters of the ultrasonic transducer to compensate for fluctuations in the cavitation field caused by changes in the state of the mixture, thus maintaining the cavitation effect of the ultrasound on the mixture in a stable state.

[0028] In specific implementation, please refer to Figure 2The homogeneous mixture, after ultrasonic-assisted dispersion, is placed in a viscosity measuring device. The device uses microtube rheology to obtain the current viscosity of the homogeneous mixture in real time. The current viscosity is a physical quantity reflecting the solute concentration and the degree of intermolecular interaction in the homogeneous mixture. The data processing device has a pre-established sedimentation medium library, which stores the identifiers, applicable viscosity ranges, and physical property parameters of several accelerated sedimentation media in a relational table format. The accelerated sedimentation media are magnetic nanobeads with hydrophilic groups on their surface. These hydrophilic groups are composed of carboxyl groups or polyethylene glycol segments, used to enhance the dispersion stability of the magnetic nanobeads in biological sample media and reduce non-specific adsorption. The data processing device compares the measured current viscosity value with the applicable viscosity range corresponding to each accelerated sedimentation medium in the sedimentation medium library, selecting the accelerated sedimentation medium whose applicable viscosity range covers the current viscosity value as the accelerated sedimentation medium to be used.

[0029] In practice, the amount of accelerating sedimentation medium added is dynamically calculated based on the total volume of the homogeneous mixture and the real-time measured viscosity. The total volume of the homogeneous mixture is recorded and stored by the sampling system before proceeding to this step. The formula for dynamically calculating the amount added is:

[0030]

[0031] in, This indicates the mass of the accelerating sedimentation medium added, expressed in milligrams. The mass concentration conversion factor ranges from 0.8 mg / mL to 1.5 mg / mL and is calibrated based on the measured relationship curve between the solid content of the nanomagnetic beads and the sedimentation efficiency. This indicates the total volume of a homogeneous mixture, expressed in milliliters. This indicates the measured value of the current viscosity, expressed in millipascal-seconds. This represents the reference viscosity value, set as the viscosity of pure water at 25 degrees Celsius, and is 0.890 millipascals per second. The viscosity contribution index is 0.6, determined by preparing simulated samples of nuclear matrix protein 22 with different viscosity gradients and measuring the viscosity at different... The coefficient of variation of the time required to achieve the same degree of sedimentation after adding nano-magnetic beads is selected, and the coefficient of variation that minimizes this coefficient is chosen. The value is used as a fixed parameter. The average particle size of the magnetic nanobeads used was determined by dynamic light scattering. The average particle size was controlled between 50 nm and 200 nm. When the average particle size was less than 50 nm, the magnetic responsiveness was insufficient, which led to a weakening of the sedimentation driving force. When the average particle size was greater than 200 nm, Brownian motion was inhibited, which was not conducive to the uniform binding with immune complexes.

[0032] In practice, after selecting the accelerated sedimentation medium and calculating its required mass, a precision syringe pump is used to add the accelerated sedimentation medium suspension dropwise to the homogeneous mixture at a preset dropping rate. The preset dropping rate is set based on the total volume of the homogeneous mixture: 10 μL / min for a total volume less than 2 mL, and 20 μL / min for a total volume greater than or equal to 2 mL. During the addition, a magnetic stir bar located at the bottom of the reaction vessel rotates continuously at 100 rpm, ensuring the accelerated sedimentation medium is uniformly dispersed in the homogeneous mixture, forming a uniformly distributed nano-magnetic bead suspension system. After the addition is complete, stirring is immediately stopped, and the reaction vessel is transferred to a settling support for settling.

[0033] In practice, during the settling stage, transmitted light signals of the homogeneous mixture along the vertical optical path are collected at fixed time intervals. The fixed time interval is set to 0.5 seconds. The light source module emits a parallel beam with a wavelength of 660 nanometers, which passes perpendicularly through the transparent optical window of the reaction vessel. The transmitted light signal is synchronously received by a photodetector on the opposite side and converted into a voltage signal. The data processing device converts the voltage signal at each time point into a turbidity value according to the Lambert-Beer law, following the acquisition sequence, and associates and stores the acquisition time with the corresponding turbidity value, thereby obtaining a series of dynamic turbidity values ​​distributed according to the acquisition time.

[0034] In practice, the data processing device reads multiple dynamic turbidity values ​​and the corresponding acquisition time for each value from the memory. Using the acquisition time as the x-axis and the dynamic turbidity value as the y-axis, a continuous and smooth turbidity decay curve is plotted on a coordinate system. Cubic spline interpolation is used to fit the discrete data points to the curve to eliminate random noise. On the turbidity decay curve, the data processing device identifies the inflection point where the turbidity value transitions from gradually increasing to gradually decreasing by observing the change in the sign of the first derivative. Specifically, the device calculates the numerical first derivative of the curve point by point along the increasing time direction. Between adjacent data points where the first derivative changes from positive to negative (crossing zero), there is a peak point. Linear interpolation is used to determine the turbidity value and time corresponding to this inflection point. The turbidity value corresponding to the inflection point is defined as the peak turbidity value, and the time corresponding to the inflection point is defined as the peak time point.

[0035] In practice, a mapping table of peak time points and concentrations of standard samples is pre-stored in the memory of the data processing device. The mapping table is constructed as follows: a series of standard solutions with known concentrations of nuclear matrix protein 22 are prepared, ranging from 0 nmol / mL to 500 nmol / mL. For each standard solution, the exact same procedure as for the test sample is performed, including monitoring turbidity in the immune response, ultrasonic-assisted dispersion during abnormal reaction periods, adding accelerated sedimentation medium, collecting dynamic turbidity values, plotting turbidity decay curves, and extracting the peak time points of the standard samples. The standard concentrations and peak time points corresponding to each standard solution are then paired and recorded to form the mapping table. Each record in the mapping table contains two fields: the peak time point and the standard concentration. After obtaining the peak time point of the sample to be tested, the data processing device calculates the numerical difference between the peak time point and the peak time points of all standards in the reference mapping table, finds the standard peak time point with the smallest absolute difference as the closest standard peak time point, reads the standard concentration corresponding to the closest standard peak time point from the reference mapping table, and outputs this standard concentration as the initial concentration of nuclear matrix protein 22.

[0036] In practice, the nonlinear correction model is stored in the permanent memory of the data processing device. Internally, the nonlinear correction model maintains a historical correction record library containing multiple historical correction records. Each historical correction record consists of two numerical fields: the first field is the historical initial concentration, calculated before the nonlinear correction step after performing a complete detection procedure on a specific nuclear matrix protein 22 sample; the second field is the historical true concentration, determined by a certified reference method for the same nuclear matrix protein 22 sample. The reference method used is enzyme-linked immunosorbent assay (ELISA), and the ELISA results, calibrated with standard materials, are recorded as the historical true concentration.

[0037] In practice, after obtaining the initial concentration of the sample to be tested through the aforementioned steps, the data processing device transmits the initial concentration numerically to the nonlinear correction model. The nonlinear correction model reads each historical correction record from the historical correction record library one by one, extracting the historical initial concentration field from each record. The model calculates the concentration difference between the initial concentration and the historical initial concentration in each historical correction record, using the absolute value of the difference between the initial concentration and the historical initial concentration. The model reads a preset difference threshold from a register, which is fixed at 5.0 nanograms per milliliter. This threshold is set based on the allowable analytical error range of nuclear matrix protein 22 at the clinical decision threshold, ensuring that the selected matching correction records are comparable to the sample to be tested in terms of concentration. The model compares the concentration difference with the preset difference threshold, selecting all historical correction records with concentration differences less than the threshold as the set of matching correction records.

[0038] In practice, the nonlinear correction model uses a weighted average of the historical true concentrations in the matched correction record set. The formula for calculating the weighted average is as follows:

[0039]

[0040] in, This indicates the final concentration of nuclear matrix protein 22; This indicates the total number of matching and correcting records in the matching and correcting record set; Indicates the first matching correction record in the set. The historical true concentration of each matching correction record; Indicates the first matching correction record in the set. The weighting coefficients of the matching correction records. According to the The concentration difference corresponding to each matching correction record is determined. The calculation method is the reciprocal of the concentration difference, that is... , Indicates the initial concentration and the first The concentration difference between the historical initial concentrations of the matched calibration records is calculated. When the concentration difference approaches zero, it is forced to be set to a very small positive number to prevent division by zero errors. The very small positive number is 0.001 nanograms per milliliter. After the weighted average calculation is completed, the nonlinear calibration model outputs the calculation result as the final concentration and writes the final concentration along with the initial concentration of this detection into the historical calibration record library, forming a new historical calibration record for subsequent detections.

[0041] In practice, the training process of the nonlinear correction model is independent of the online computation process and is completed before model deployment. When training the nonlinear correction model, several known true concentrations of nuclear matrix protein 22 standards are collected, with no fewer than 200 standards. The concentrations of nuclear matrix protein 22 in the standards cover the range from 0 nanograms per milliliter to 500 nanograms per milliliter, and the distribution within each concentration range satisfies the uniform sampling principle. For each standard, the initial training concentration corresponding to the standard is obtained using the exact same detection procedure as the test sample. The detection process includes the following steps: mixing the standard and antibody reagent to form an initial mixture; continuously monitoring multiple initial turbidity values ​​of the initial mixture within a preset reaction period; identifying abnormal reaction periods of the initial mixture based on the multiple initial turbidity values, and performing ultrasonic-assisted dispersion treatment on the initial mixture during these abnormal periods to obtain a homogeneous mixture; adding an accelerating sedimentation medium to the homogeneous mixture and acquiring multiple dynamic turbidity values ​​of the homogeneous mixture in real time during the sedimentation process; extracting the peak turbidity value and peak time point from the multiple dynamic turbidity values, and calculating the initial training concentration based on the peak time point and the mapping table between the peak time point and the standard concentration. The initial training concentration of each standard and its corresponding known true concentration are combined to form a training data pair, and all training data pairs corresponding to all standards constitute the training dataset.

[0042] In the specific implementation, a feedforward neural network is constructed as the model framework to be trained. The structure of the feedforward neural network is as follows: the input layer contains one node, used to input the initial concentration for training; after the input layer, a first hidden layer is connected, containing 16 neurons with a linear rectified function as the activation function; after the first hidden layer, a second hidden layer is connected, containing 8 neurons with a linear rectified function as the activation function; after the second hidden layer, an output layer is connected, containing one neuron, with no activation function set, directly outputting the fitted concentration value. All layers of the feedforward neural network are fully connected, and the initial values ​​of the inter-layer weights are generated using the Xavier uniform distribution initialization method, while the initial value of the bias term is set to zero.

[0043] In practice, the initial concentrations in the training dataset are used as input features to the feedforward neural network, and the corresponding known true concentrations in the training dataset are used as output labels. The feedforward neural network is trained using the backpropagation algorithm. During training, the mean squared error function is used as the loss function, and the adaptive moment estimation optimizer is selected. The learning rate of the adaptive moment estimation optimizer is set to 0.001, the first-order moment decay factor is set to 0.9, and the second-order moment decay factor is set to 0.999. For each training iteration, 32 training data pairs are randomly selected from the training dataset as a batch and input into the feedforward neural network for forward propagation. After calculating the loss function value, the weights and biases are updated through backpropagation. Each iteration through all training data pairs is considered one epoch. After each epoch, the average of the loss function values ​​of all batches within that epoch is calculated as the average loss for that epoch. When the average loss of a epoch drops below a preset loss threshold, training stops, and the parameters of the feedforward neural network at this point are fixed and exported for deployment as a nonlinear correction model. The preset loss threshold is set to 0.01. The basis for this preset loss threshold is that when the mean squared error loss value is less than 0.01, the deviation between the output concentration of the nonlinear correction model and the known true concentration is less than the clinically acceptable concentration deviation limit, and the model has practical usability.

[0044] In specific implementation, please refer to Figure 3 The initial mixture of the test sample and antibody reagent is collected, and multiple initial turbidity values ​​of the initial mixture are continuously monitored within a preset reaction period. This step is accomplished collaboratively by a sample loading module, a thermostatic stirring module, and an optical detection module integrated into the automated analyzer. The sample loading module is equipped with a high-precision pipette pump and disposable pipette tips, with a minimum aspiration step of 0.1 μL. The thermostatic stirring module incorporates a Peltier temperature control element and a DC brushless stirring motor, while the optical detection module consists of a monochromatic light source, a collimating lens, a cuvette holder, and a silicon photodiode.

[0045] In practice, the sample to be tested is injected into a reaction vessel containing antibody reagent according to a preset sample volume to reagent volume ratio. The preset sample volume to reagent volume ratio is set to 1:19, meaning that 1 volume unit of sample to be tested corresponds to 19 volume units of antibody reagent. This volume ratio is set based on the following: the molar amount of anti-matrixonal protein 22 monoclonal antibody in the antibody reagent needs to be at least 50 times excess relative to the maximum possible molar amount of matrixonal protein 22 in the sample to ensure that all target antigens are completely bound by the antibody; when the sample volume is 10 μL, the antibody reagent volume is 190 μL, and the total reaction volume is 200 μL. This total volume can completely fill the illumination area of ​​the optical detection path, avoiding optical path distortion caused by the curvature of the liquid surface. The sample loading module first aspirates 10 μL of sample to be tested using a pipette pump, then moves it above the reaction vessel and injects the sample into the reaction vessel, which has already been pre-filled with 190 μL of antibody reagent by the reagent dispensing unit.

[0046] In practice, after the sample to be tested is injected, the reaction vessel is stirred at a first rotation speed for a first preset time under constant temperature conditions to obtain the initial mixture. The constant temperature condition is set so that the ambient air temperature around the reaction vessel is maintained at 37 degrees Celsius, and the Peltier temperature control element uses a proportional-integral-differential algorithm to control temperature fluctuations within ±0.1 degrees Celsius. The first rotation speed is set to 600 revolutions per minute, and the first preset time is set to 15 seconds. The reason for setting the first rotation speed to 600 revolutions per minute is that when the rotation speed is lower than 400 revolutions per minute, high-concentration samples may cause temporary antigen excess in the local area of ​​sample addition, resulting in uneven immunoprecipitation. When the rotation speed is higher than 800 revolutions per minute, the violent agitation of the liquid surface can easily introduce air bubbles and interfere with subsequent turbidity measurements. Selecting 600 revolutions per minute can achieve a turbulent mixing state with a Reynolds number of approximately 800 within 15 seconds, ensuring that the sample to be tested and the antibody reagent are in full contact and the immune reaction is initiated. After stirring for 15 seconds, the stirring motor immediately stops running, and the mixture at this time is the initial mixture.

[0047] In practice, within a preset reaction period, a light source is controlled to vertically irradiate the initial mixture at a fixed intensity, and a photodetector receives the transmitted light signal at a fixed sampling frequency. The preset reaction period is defined as from the moment stirring stops until 300 seconds later, covering the main stages of immune complex formation and primary aggregation. A light-emitting diode (LED) with a center wavelength of 670 nm is used as the light source, and its driving current is kept constant at 20 mA, maintaining a fixed light intensity of 18,000 millicandela. The light beam emitted by the LED is collimated by a collimating lens to form a parallel beam with a diameter of 2 mm, which passes perpendicularly through the optical measurement area of ​​the reaction cup. On the opposite side of the reaction cup, a silicon photodiode, with a photosensitive area of ​​7.6 square millimeters, acts as a photodetector to receive the transmitted light signal. The photodetector collects the transmitted light signal at a fixed sampling frequency of 20 Hz, meaning the transmitted light intensity is recorded every 50 milliseconds.

[0048] In practice, the transmitted light signal received at each sampling moment is converted into a corresponding turbidity value, resulting in multiple initial turbidity values. The photodetector converts the received transmitted light intensity into a corresponding analog voltage signal, which is then converted into a digital signal by a 12-bit analog-to-digital converter. The digital signal value is denoted as... The data processing device reads the reference signal of transmitted light when the blank cuvette is filled with pure water. The transmitted light reference signal Measurements are taken and stored during each power-on calibration. The data processing unit uses... and Based on the turbidity conversion formula

[0049]

[0050] in, This represents the calculated turbidity value, expressed in units of per centimeter. The optical path length of the reaction cup is fixed at 1 cm. This represents the digital signal value output by the photodetector at the current sampling time; The transmitted light reference signal is a digital signal value output by the photodetector under the condition that the reaction cup is filled with pure water and the light source is stable. It represents the intensity of the full transmission signal when there is no turbidity attenuation. The dark signal value is the digital signal output by the photodetector when the light source is completely blocked. It represents the contribution of circuit background noise and dark current. The data is collected and stored before each measurement by closing the optical path using a mechanical shutter. The data processing device calculates the turbidity value obtained at each sampling moment within the preset reaction time period. The timestamp corresponding to the sampling time is stored to form multiple initial turbidity values.

[0051] In practice, after obtaining the urine sample to be tested, the urine sample is transferred to a centrifuge tube. The urine sample is the midstream morning urine excreted by the subject, collected in a clean disposable urine cup. 10 ml of the urine sample is then drawn using a pipette and injected into a 15 ml conical-bottom centrifuge tube. The centrifuge tube is made of medical-grade polypropylene and has a maximum centrifugal force tolerance of 12000g.

[0052] In practice, centrifuge tubes containing urine samples are placed into the centrifuge rotor. The centrifuge is a benchtop high-speed refrigerated centrifuge with a horizontal rotor, capable of holding six centrifuge tubes. The centrifuge is controlled to centrifuge the urine samples in the tubes at a second speed for a second preset time. The second speed is set to 1500g, and the second preset time is set to 10 minutes. The second centrifugal speed setting of 1500g is based on the fact that common precipitates in urine samples, such as cell debris, magnesium ammonium phosphate crystals, and urate crystals, can be completely settled within 10 minutes under a centrifugal force of 900g or higher. A centrifugal force of 1500g ensures sufficient compaction of the precipitate while avoiding excessive centrifugation that could break down large protein molecules or mucoproteins. An evaluation of the consistency between the supernatant turbidity and total protein content measurements after centrifuging 200 clinical urine samples at three different centrifugal forces (1000g, 1500g, and 2000g) showed that the intra-batch coefficient of variation for supernatant turbidity was lowest at 1500g, and the total protein content deviated by no more than 3% from that at 1000g. During centrifugation, the centrifuge chamber temperature was maintained at 4 degrees Celsius to inhibit bacterial growth and the degradation of nuclear matrix protein 22 by proteases. After centrifugation reached the second preset time, the centrifuge slowed down and stopped. Upon removing the centrifuge tubes, a dense precipitate was observed at the bottom, with a clear supernatant above. Using a pipette, slowly aspirate the supernatant from 3 mm below the liquid surface in the centrifuge tube and transfer it to a clean sample storage tube. Avoid touching or disturbing the bottom precipitate during the aspiration process. This supernatant is the supernatant to be used for total protein measurement.

[0053] In practice, the total protein content of the supernatant was measured using a dicaprinic acid protein quantification kit. 20 μL of supernatant was added to each well of a microplate, followed by 200 μL of chromogenic working solution. This working solution was freshly prepared from dicaprinic acid solution, copper sulfate solution, and alkaline buffer at a volume ratio of 50:1:50. The microplate was placed in a constant-temperature shaker and incubated at 37°C for 30 minutes. After incubation, the microplate was removed, and the absorbance at 562 nm was read using a microplate spectrophotometer. Based on the absorbance-protein concentration standard curve prepared using bovine serum albumin standards, the absorbance value was converted to total protein content, expressed in micrograms per milliliter (µg / mL). The data processing unit compared the measured total protein content with a pre-stored protein threshold in a register. The protein threshold was set to 100 µg / mL. The protein threshold of 100 micrograms per milliliter is set based on the following: the optimal antigen concentration range for antigen-antibody reaction in immunoturbidimetry is from nanograms to low micrograms per milliliter. When the total protein concentration of a sample exceeds 100 micrograms per milliliter, the interference of background proteins on immune complex formation and turbidity signals increases significantly. Preliminary experimental data show that in urine samples with a total protein concentration exceeding 100 micrograms per milliliter, the linear correlation coefficient between the initial turbidity signal and the true concentration of nuclear matrix protein 22 drops below 0.85, severely affecting the accuracy of detection. When the total protein content is below 100 micrograms per milliliter, the supernatant is considered a qualified sample and its information is recorded as a sample to be tested. When the total protein content reaches or exceeds 100 micrograms per milliliter, the data processing device issues a sample dilution prompt. The operator takes a portion of the supernatant and dilutes it 2-fold with phosphate buffer. After dilution, the total protein content is measured again until the total protein content drops below the protein threshold. The diluted supernatant is then used as the sample to be tested, and the dilution factor is recorded for subsequent concentration conversion.

[0054] In practice, after the nonlinear correction model outputs the final concentration of matrix nuclear protein 22, the data processing device accesses the patient history test record database in the laboratory information management system via a network interface. Using the unique identifier of the patient to whom the current sample belongs as the query key, the data processing device retrieves the patient's most recent matrix nuclear protein 22 test record from the patient history test record database and reads the patient history test concentration of matrix nuclear protein 22 from this record. The patient history test concentration represents the final concentration value output by the nonlinear correction model in the previous test. If no historical test record for the patient is found in the patient history test record database, the retesting process is skipped, and the current final concentration is directly archived as the report result.

[0055] In practice, after retrieving the patient's historical concentration readings, the data processing device calculates the rate of change between the final concentration and the patient's historical concentration readings. The formula for calculating the rate of change is:

[0056]

[0057] in, The rate of change in concentration is expressed as a percentage. This represents the final concentration of nuclear matrix protein 22 output by the nonlinear correction model, in nanograms per milliliter. This indicates the patient-historical concentration of nuclear matrix protein 22 retrieved from the patient history test record database, in nanograms per milliliter. (The molecule contains...) This represents the absolute value of the concentration difference between the final concentration and the patient's historical concentration.

[0058] In practical implementation, the data processing device will calculate the concentration change rate. The result was compared with a preset change rate threshold. The preset change rate threshold was set at 35%. The rationale for setting the preset change rate threshold at 35% is that, as a post-operative monitoring biomarker for bladder cancer, the concentration change of nuclear matrix protein 22, after combining physiological fluctuations with the coefficient of variation of the detection method, typically does not exceed 30% in two consecutive short-term measurements without clinical progression. A 35% threshold allows for approximately 5% analytical margin, avoiding frequent and unnecessary retesting. When the concentration change rate... If the concentration change is less than 35%, the final concentration is considered to be within the allowable fluctuation range compared to the patient's historical test concentrations. The final concentration will then be output as the test report value and archived. (When the concentration change rate...) When the concentration change rate is greater than or equal to 35%, it exceeds the preset change rate threshold, indicating that the current test result deviates significantly from the historical level, and random errors in sampling or operation during the testing process need to be excluded.

[0059] In practice, when the concentration change rate exceeds a preset threshold, the data processing device generates a retest marker. This marker is a digital identifier containing the sample number, original final concentration, patient's historical test concentrations, and concentration change rate. The retest marker is written to the test log storage. Simultaneously, the data processing device sends a retest command to the autosampler control unit. Upon receiving the retest command, the autosampler control unit moves the autosampler arm above the sample storage tube containing the supernatant of the sample to be tested. The autosampler's aspiration needle then inserts into the storage tube to extract a new retest sample from the retained supernatant. The volume of the retest sample is 10 microliters.

[0060] In practice, the retested samples are retested using the same calculation process as the final concentration. This same calculation process includes: mixing the retested samples with antibody reagents to form an initial mixture; continuously monitoring multiple initial turbidity values ​​of the initial mixture within a preset reaction period; identifying abnormal reaction periods based on these initial turbidity values ​​and performing ultrasonic-assisted dispersion during these abnormal periods to obtain a homogeneous mixture; adding an accelerating sedimentation medium to the homogeneous mixture and acquiring multiple dynamic turbidity values ​​during the sedimentation process in real time; extracting the peak turbidity value and peak time point from these dynamic turbidity values; calculating the initial retest concentration of nuclear matrix protein 22 based on a mapping table of peak time points and standard concentrations; inputting the initial retest concentration into a nonlinear correction model; and outputting the retest concentration from the nonlinear correction model. The data processing device recalculates the concentration change rate by comparing the retested concentration with the patient's historical test concentration. If the recalculated concentration change rate still exceeds the preset change rate threshold, the value that is closer to the patient's historical test concentration between the retested concentration and the original final concentration is used as the test report value. If the recalculated concentration change rate is lower than the preset change rate threshold, the retested concentration replaces the original final concentration, and the retested concentration is output as the test report value of the sample to be tested and archived.

[0061] The foregoing has provided a detailed description of one embodiment of the present invention, but this description is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the claims of this invention should still fall within the patent coverage of this invention.

Claims

1. A method for detecting and analyzing nuclear matrix protein 22 based on immunoturbidimetry, characterized in that, The method includes: An initial mixture of the sample to be tested and the antibody reagent is collected, and multiple initial turbidity values ​​of the initial mixture are continuously monitored within a preset reaction period. Based on the multiple initial turbidity values, the abnormal reaction period of the initial mixture is identified, and during the abnormal reaction period, the initial mixture is subjected to ultrasonic-assisted dispersion treatment to obtain a homogeneous mixture; An accelerating sedimentation medium was added to the homogeneous mixture, and multiple dynamic turbidity values ​​of the homogeneous mixture were collected in real time during the sedimentation process. The peak turbidity value and its corresponding peak time point are extracted from the plurality of dynamic turbidity values, and the initial concentration of nuclear matrix protein 22 is calculated based on the peak turbidity value and the peak time point, including: The multiple dynamic turbidity values ​​were plotted as turbidity decay curves in chronological order of collection time. Identify the inflection point in the turbidity decay curve where the turbidity value changes from rising to falling, determine the turbidity value at the inflection point as the peak turbidity value, and determine the time point corresponding to the inflection point as the peak time point; Obtain a pre-established mapping table of peak time points and concentrations of standard substances; The peak time point of the standard closest to the peak time point is found in the reference mapping table, and the corresponding standard concentration is read as the initial concentration of the nuclear matrix protein 22. The initial concentration is input into a pre-trained nonlinear correction model, which outputs the final concentration of nuclear matrix protein 22.

2. The method for detecting and analyzing nuclear matrix protein 22 based on immunoturbidimetry according to claim 1, characterized in that, The step of identifying the abnormal reaction period of the initial mixture based on the plurality of initial turbidity values ​​includes: The multiple initial turbidity values ​​are arranged into a turbidity time series in chronological order. Calculate the difference between two adjacent initial turbidity values ​​in the turbidity time series to obtain multiple turbidity change rates; Filter out abnormal turbidity change rates whose absolute values ​​exceed a preset change rate threshold; The collection period corresponding to the abnormal rate of change is marked as the abnormal reaction period.

3. The method for detecting and analyzing nuclear matrix protein 22 based on immunoturbidimetry according to claim 2, characterized in that, The initial mixture is subjected to ultrasonic-assisted dispersion treatment during the abnormal reaction period to obtain a homogeneous mixture, comprising: Determine the maximum turbidity fluctuation amplitude that occurs in the initial mixture during the abnormal reaction period; The initial operating frequency and initial operating power of the ultrasonic transducer are selected based on the maximum turbidity fluctuation amplitude. During the abnormal reaction period, the ultrasonic transducer is controlled to emit ultrasonic waves into the initial mixture according to the initial operating frequency and the initial operating power. The real-time turbidity fluctuation amplitude of the initial mixture during the ultrasonic treatment process is monitored. When the real-time turbidity fluctuation amplitude drops below a preset fluctuation threshold, the ultrasonic waves are stopped, and the homogeneous mixture is obtained.

4. The method for detecting and analyzing nuclear matrix protein 22 based on immunoturbidimetry according to claim 3, characterized in that, The ultrasonic transducer simultaneously acquires ultrasonic echo signals while emitting ultrasonic waves, and adjusts the initial operating frequency and the initial operating power in real time based on the ultrasonic echo signals.

5. The method for detecting and analyzing nuclear matrix protein 22 based on immunoturbidimetry according to claim 1, characterized in that, An accelerating sedimentation medium is added to the homogeneous mixture, and multiple dynamic turbidity values ​​of the homogeneous mixture are collected in real time during the sedimentation process, including: Select an accelerated sedimentation medium from a pre-stored sedimentation medium library that matches the current viscosity of the homogeneous mixture. The accelerated sedimentation medium is a nanomagnetic bead with a surface modified with hydrophilic groups. The selected accelerating sedimentation medium is added to the homogeneous mixture at a preset dropping rate, and the homogeneous mixture is continuously stirred during the dropping process. During the settling phase after the addition of the solution, multiple dynamic turbidity values ​​of the homogeneous mixture on the vertical optical path are collected at fixed time intervals.

6. The method for detecting and analyzing nuclear matrix protein 22 based on immunoturbidimetry according to claim 5, characterized in that, The amount of the accelerating sedimentation medium added is determined dynamically based on the volume of the homogeneous mixture and the current viscosity, and the average particle size of the nanomagnetic beads is between 50nm and 200nm.

7. The method for detecting and analyzing nuclear matrix protein 22 based on immunoturbidimetry according to claim 1, characterized in that, The initial concentration is input into a pre-trained nonlinear correction model, which outputs the final concentration of nuclear matrix protein 22, including: Obtain multiple historical correction records stored in the nonlinear correction model, each of which contains the historical initial concentration and the corresponding historical true concentration; Calculate the concentration difference between the initial concentration and the historical initial concentration in each historical correction record; Select matching correction records where the concentration difference is less than a preset difference threshold; The historical true concentrations in the matching correction records are weighted and averaged, and the weighted average result is used as the final concentration.

8. The method for detecting and analyzing nuclear matrix protein 22 based on immunoturbidimetry according to claim 7, characterized in that, The training process of the nonlinear correction model includes: Collect multiple nuclear matrix protein 22 standards with known real concentrations, and use the same calculation method as the initial concentration to obtain the training initial concentration corresponding to each standard; The initial training concentration and the corresponding actual concentration of each of the aforementioned standards are used to form training data pairs; Using the initial training concentration as the input feature and the true concentration as the output label, a backpropagation algorithm is used to train a feedforward neural network until the loss function value of the feedforward neural network converges to a preset loss threshold. The trained feedforward neural network is used as the nonlinear correction model.

9. The method for detecting and analyzing nuclear matrix protein 22 based on immunoturbidimetry according to claim 1, characterized in that, The process of collecting the initial mixture of the test sample and antibody reagent, and continuously monitoring multiple initial turbidity values ​​of the initial mixture within a preset reaction period, includes: According to the preset sample volume to reagent volume ratio, the sample to be tested is injected into the reaction cup containing the antibody reagent; The reaction vessel is stirred at a first rotation speed for a first preset time under constant temperature conditions to obtain the initial mixture. During the preset reaction period, the light source is controlled to vertically irradiate the initial mixture with a fixed light intensity, and the transmitted light signal is received by a photodetector at a fixed sampling frequency. The transmitted light signal received at each sampling time is converted into a corresponding turbidity value to obtain the plurality of initial turbidity values.

Images