Biomarker detection method and system for systemic lupus erythematosus

Abstract

The present application relates to the technical field of biomarker detection data analysis, and discloses a biomarker detection method and system for systemic lupus erythematosus. The method comprises: collecting the detection data of anti-dsDNA antibodies, anti-Sm antibodies and anti-nRNP antibodies through the detection data acquisition module connected to the biomarker detection device to obtain the original detection data set; performing standardization calculation on the data set to obtain a standardized data matrix; calculating the correlation coefficient and distribution characteristics to obtain the associated feature vector; analyzing the dynamic change characteristics to obtain an evaluation report; performing feature pattern recognition to obtain early warning information; and finally obtaining the detection data analysis results. The present application can process multiple biomarker detection data at the same time, and analyze their correlation and change trends, thereby realizing comprehensive processing and analysis of the detection data.

Description

Technical Field

[0001] The present application relates to the field of biomarker detection data analysis, and in particular to a biomarker detection method and system for systemic lupus erythematosus. Background Art

[0002] The detection of biomarkers related to systemic lupus erythematosus (SLE) is an important data collection task, among which anti-double-stranded (ds) DNA antibodies, anti-Sm antibodies and anti-nRNP antibodies are important biomarkers. The current detection methods mainly use enzyme-linked immunosorbent assay (ELISA), indirect immunofluorescence (IIF) and other technical means to obtain the detection data of these antibodies. In the existing technology, the processing of detection data mainly includes numerical acquisition, data standardization and basic statistical analysis. These data are directly output by the detection equipment to form basic data records.

[0003] However, existing biomarker detection data processing methods have obvious shortcomings. First, there is a lack of a unified standardized processing process for the detection data of multiple biomarkers, which makes it difficult to directly compare data between different detection batches. Secondly, existing methods mainly focus on the numerical recording of a single marker, ignoring the relationship between multiple marker detection data. Especially in terms of data processing, existing methods are often only able to perform simple data recording and basic statistics, and lack the ability to deeply analyze data change trends. These deficiencies have seriously affected the analysis effect of biomarker detection data and the mining of data value. Summary of the invention

[0004] The present application provides a biomarker detection method and system for systemic lupus erythematosus, which can simultaneously process multiple biomarker detection data and analyze their correlation and change trends, thereby achieving comprehensive processing and analysis of the detection data.

[0005] In a first aspect, the present application provides a biomarker detection method for systemic lupus erythematosus, and the biomarker detection method for systemic lupus erythematosus includes: connecting a biomarker detection device through a detection data acquisition module, and collecting the detection data of anti-dsDNA antibodies, anti-Sm antibodies and anti-nRNP antibodies in real time to obtain a biomarker original detection data set; performing unit uniform conversion and numerical standardization calculation on the biomarker original detection data set to obtain a standardized biomarker data matrix; based on the standardized biomarker data matrix, using a multidimensional data association analysis algorithm to calculate the correlation coefficient and numerical distribution characteristics between the three antibody detection values ​​to obtain a biomarker association feature vector; based on the biomarker association feature vector, analyzing the dynamic change characteristics of the detection value to obtain a biomarker detection evaluation report; according to the biomarker detection evaluation report, performing feature pattern recognition on abnormal changes in the detection value to obtain detection data warning information; based on the detection data warning information, performing dynamic feature analysis on the biomarker detection data to obtain a biomarker detection data analysis result.

[0006] In a second aspect, the present application provides a biomarker detection system for systemic lupus erythematosus, the biomarker detection system for systemic lupus erythematosus comprising:

[0007] The acquisition module is used to connect the biomarker detection device through the detection data acquisition module, collect the detection data of anti-dsDNA antibodies, anti-Sm antibodies and anti-nRNP antibodies in real time, and obtain the original detection data set of the biomarker;

[0008] A conversion module, used to perform unit uniform conversion and numerical standardization calculation on the original biomarker detection data set to obtain a standardized biomarker data matrix;

[0009] A calculation module, for calculating the correlation coefficient and numerical distribution characteristics between the three antibody detection values ​​based on the standardized biomarker data matrix using a multidimensional data association analysis algorithm to obtain a biomarker association feature vector;

[0010] An evaluation module, used to analyze the dynamic change characteristics of the detection value according to the biomarker-associated feature vector to obtain a biomarker detection evaluation report;

[0011] An identification module, used to perform feature pattern recognition on abnormal changes in detection values ​​according to the biomarker detection evaluation report to obtain detection data warning information;

[0012] The analysis module is used to perform dynamic feature analysis on the biomarker detection data based on the detection data warning information to obtain the biomarker detection data analysis results.

[0013] In the technical solution provided by the present application, real-time data collection of biomarker detection equipment is realized, ensuring the timely acquisition and integrity of the detection data of three biomarkers, namely, anti-dsDNA antibodies, anti-Sm antibodies and anti-nRNP antibodies. By uniformly converting the units of the original detection data set of biomarkers and calculating the numerical standardization, the comparability problem of data under different batches and different detection conditions is solved, and the standardized biomarker data matrix is ​​obtained to lay the foundation for subsequent analysis. The correlation coefficient and numerical distribution characteristics between the detection values ​​of the three antibodies are calculated using the multidimensional data association analysis algorithm, which not only realizes the joint analysis of multiple biomarker data, but also automatically extracts the association characteristics between the data through the algorithm, and the obtained biomarker association feature vector can fully reflect the distribution law of the data. The dynamic change characteristics of the detection values ​​are analyzed, and a biomarker detection evaluation report is generated, which realizes the quantitative description of the data change trend. The abnormal changes of the detection values ​​are analyzed by feature pattern recognition technology, and a data-based early warning mechanism is established, and the obtained detection data early warning information is objective and reliable. Finally, through dynamic feature analysis, the biomarker detection data was comprehensively analyzed, and the analysis results obtained covered multiple dimensional characteristics of the data, which not only improved the automation level of data processing, but also enhanced the depth and accuracy of data analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0015] Figure 1 This is a schematic diagram of an embodiment of a biomarker detection method for systemic lupus erythematosus in the embodiments of the present application;

[0016] Figure 2 This is a flow chart of the steps of performing unified unit conversion and numerical standardization calculation on the original detection data set of biomarkers in the embodiments of the present application;

[0017] Figure 3 A timing diagram of the steps of calculating the correlation coefficient and numerical distribution characteristics between the detection values ​​of three antibodies in the embodiment of the present application;

[0018] Figure 4 This is a schematic diagram of an embodiment of a biomarker detection system for systemic lupus erythematosus in the embodiments of the present application. DETAILED DESCRIPTION

[0019] The embodiments of the present application provide a method and system for detecting biomarkers for systemic lupus erythematosus. The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments described here can be implemented in an order other than that illustrated or described here. In addition, the terms "including" or "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0020] For ease of understanding, the specific process of the embodiment of the present application is described below. Figure 1 In the embodiments of the present application, one embodiment of the biomarker detection method for systemic lupus erythematosus includes:

[0021] Step S101, connecting the biomarker detection device through the detection data acquisition module, and collecting the detection data of anti-dsDNA antibodies, anti-Sm antibodies and anti-nRNP antibodies in real time to obtain the original detection data set of the biomarker;

[0022] Step S102, performing unit uniform conversion and numerical standardization calculation on the original biomarker detection data set to obtain a standardized biomarker data matrix;

[0023] Step S103, based on the standardized biomarker data matrix, using a multidimensional data association analysis algorithm to calculate the correlation coefficient and numerical distribution characteristics between the three antibody detection values, to obtain a biomarker association feature vector;

[0024] Step S104: Analyze the dynamic change characteristics of the detection value according to the biomarker-related feature vector to obtain a biomarker detection evaluation report;

[0025] Step S105: According to the biomarker detection evaluation report, feature pattern recognition is performed on abnormal changes in the detection values ​​to obtain detection data warning information;

[0026] Step S106: Based on the detection data warning information, dynamic feature analysis is performed on the biomarker detection data to obtain a biomarker detection data analysis result.

[0027] It is understandable that the execution subject of the present application may be a biomarker detection system for systemic lupus erythematosus, or a terminal or a server, which is not limited here. The present application embodiment is described by taking a server as the execution subject as an example.

[0028] Specifically, the biomarker detection equipment collects the detection data of anti-dsDNA antibodies, anti-Sm antibodies and anti-nRNP antibodies, which are important biomarkers of systemic lupus erythematosus. The detection equipment detects the antibody concentration through fluorescent signals and converts the fluorescent signals into digital signals. When each sample is collected, the detection time, temperature parameters, humidity parameters, and reagent batch number information are recorded to form the original biomarker detection data set.

[0029] When the original biomarker detection data set is standardized, the anti-dsDNA antibody detection data is converted according to the international unit IU / mL and standardized by subtracting the baseline value and dividing it by the control value. For the anti-Sm antibody detection data, the signal intensity value is divided by the slope of the calibration curve for quantitative conversion. The anti-nRNP antibody detection data needs to be normalized and the values ​​are converted to the range of 0-1. The three antibody data after standardization are arranged in time series to form a standardized biomarker data matrix. The correlation between the three antibody detection values ​​is calculated using multidimensional data association analysis. The data subsets are divided according to the time window, and the correlation coefficients between anti-dsDNA antibodies and anti-Sm antibodies, anti-Sm antibodies and anti-nRNP antibodies, and anti-dsDNA antibodies and anti-nRNP antibodies are calculated. The correlation coefficient sequence is statistically analyzed, including the calculation of the mean, variance, quartiles, kurtosis coefficient and skewness coefficient, to obtain the distribution characteristic data. The correlation coefficient and distribution characteristics are combined to form a biomarker association feature vector.

[0030] Analyze dynamic change characteristics based on biomarker-associated feature vectors. Split the feature vectors by time series, calculate the rate of change in each time period, and obtain a dynamic change rate sequence. Identify the value fluctuation interval through extreme point detection, mark the rising and falling intervals, and calculate the slope value. Perform periodic analysis on the change trend, count the duration of the rising and falling intervals, calculate the correlation between the change amplitude and duration, and form a biomarker detection evaluation report. Identify abnormal change characteristics based on the biomarker detection evaluation report. Analyze the detection values ​​by time period, and calculate the mean and standard deviation as a benchmark. Mark the abnormal points that exceed the range of three times the standard deviation, and cluster the adjacent abnormal points to obtain the abnormal interval. Quantify the degree of deviation of the abnormal value, count the frequency and duration of abnormal characteristics, and generate warning information for detection data.

[0031] Finally, the dynamic characteristics of biomarker detection data are analyzed. The warning information is divided into observation windows, the warning frequency is calculated and its changing trend is analyzed. The rising and falling trends are marked, the duration of trend changes is calculated, and the occurrence patterns of different types of trends are statistically analyzed. The characteristics such as warning frequency, trend changes, duration and cycle patterns are comprehensively analyzed to form the analysis results of biomarker detection data.

[0032] For example, serum samples from patients are collected for antibody testing. The detection equipment collects antibody fluorescence signals and converts them into digital signal intensity values. The original signal value of the anti-dsDNA antibody is subtracted from the baseline value of the blank control and then divided by the control value of the standard for standardization. The Pearson correlation coefficient between the three antibody test values ​​is calculated to identify areas where the test values ​​are abnormally increased or decreased. The degree of abnormality is quantified and the duration of the abnormality is recorded. By analyzing the periodic characteristics of the early warning information, the law of changes in biomarker levels is revealed, providing data support for clinical monitoring.

[0033] In the embodiment of the present application, the real-time data collection of the biomarker detection equipment is realized, and the timely acquisition and integrity of the three biomarker detection data of anti-dsDNA antibody, anti-Sm antibody and anti-nRNP antibody are ensured. By performing unit uniform conversion and numerical standardization calculation on the original detection data set of biomarkers, the comparability problem of data under different batches and different detection conditions is solved, and the standardized biomarker data matrix is ​​obtained to lay the foundation for subsequent analysis. The correlation coefficient and numerical distribution characteristics between the detection values ​​of the three antibodies are calculated using the multidimensional data association analysis algorithm, which not only realizes the joint analysis of multiple biomarker data, but also automatically extracts the association characteristics between the data through the algorithm, and the obtained biomarker association feature vector can fully reflect the distribution law of the data. The dynamic change characteristics of the detection value are analyzed, and a biomarker detection evaluation report is generated, which realizes the quantitative description of the data change trend. The abnormal changes of the detection value are analyzed by feature pattern recognition technology, and a data-based early warning mechanism is established, and the obtained detection data early warning information is objective and reliable. Finally, the biomarker detection data is comprehensively analyzed by dynamic feature analysis, and the analysis results obtained cover the multiple dimensional characteristics of the data, which not only improves the automation level of data processing, but also enhances the depth and accuracy of data analysis.

[0034] In a specific embodiment, the process of executing step S101 may specifically include the following steps:

[0035] (1) receiving the fluorescence intensity signal of the anti-dsDNA antibody through a biomarker detection device, digitally converting the fluorescence signal, and generating first detection signal data;

[0036] (2) receiving the reaction signal of the anti-Sm antibody through the biomarker detection device, converting the intensity of the reaction signal, and generating second detection signal data;

[0037] (3) receiving the reaction signal of the anti-nRNP antibody through the biomarker detection device, converting the intensity of the reaction signal, and generating third detection signal data;

[0038] (4) Associating and marking the first detection signal data, the second detection signal data, the third detection signal data and the detection time information to generate a detection data group;

[0039] (5) Perform quality inspection on the test data set, remove abnormal data points, and obtain qualified test data;

[0040] (6) Assign a unique identification code to the qualified test data to generate the original biomarker test data set.

[0041] Specifically, the biomarker detection device receives the fluorescence intensity signal of the anti-dsDNA antibody, which is received by the photomultiplier tube in the detection instrument. The fluorescence emitted by the sample is converted into a voltage signal. The voltage signal is digitized by an analog-to-digital converter with a sampling frequency of 1kHz and a resolution of 16 bits, converting the continuous analog signal into a discrete digital signal. The converted digital signal is digitally filtered to remove high-frequency noise and form the first detection signal data.

[0042] Then the biomarker detection device receives the reaction signal of the anti-Sm antibody. After the anti-Sm antibody binds to the marker, chemiluminescence is generated. The luminescent signal is collected by the photodetector and converted into an electrical signal. The electrical signal is amplified and filtered by the signal conditioning circuit, and then the signal intensity is mapped to the standard range by the signal intensity conversion circuit. The signal intensity conversion adopts a piecewise linear mapping method to divide the original signal into multiple intervals, and each interval uses a different conversion coefficient. After conversion, the second detection signal data is obtained.

[0043] For the detection of anti-nRNP antibodies, the biomarker detection device also receives its reaction signal. The light signal generated by the reaction of anti-nRNP antibodies and substrates is photoelectrically converted to obtain a voltage signal. The voltage signal is amplified by the signal amplifier and then digitally converted by a 16-bit ADC. The digital signal is mean filtered to eliminate random noise, and then intensity converted by a signal calibration circuit to obtain the third detection signal data.

[0044] When associating the three detection signal data with the detection time information, a unified timestamp format is used, accurate to the millisecond level. Each detection data point is attached with the corresponding acquisition time information, including year, month, day, hour, minute, second, and millisecond. The time information is used as the data index, and the first detection signal data, the second detection signal data, and the third detection signal data are organized in time sequence to form a detection data group. Each group of data contains the detection values ​​of the three antibodies and their corresponding timestamps. When performing quality inspection on the detection data group, the integrity of the data is checked and data points with missing values ​​are eliminated. Then, outliers are identified by the 3σ principle, the mean and standard deviation of each signal are calculated, and data points that deviate from the mean by more than three times the standard deviation are marked as outliers. For continuously occurring outliers, it is determined whether they are caused by equipment failure. If the continuous outliers exceed the preset threshold, the entire data segment is marked as invalid. After the outliers are eliminated, qualified detection data is obtained. Finally, a unique identification code is assigned to the qualified detection data. The identification code consists of the detection date, sample number, and detection batch number, and adopts a fixed-length string format. The test date is represented by 8 digits (YYYYMMDD), the sample number is 6 digits, and the test batch number is 4 digits, and the three parts are connected by hyphens. The qualified test data are sorted and archived through the unique identification code to form the original biomarker test data set.

[0045] For example, when a serum sample is tested, the biomarker detection device collects the fluorescence signal of the anti-dsDNA antibody. The original fluorescence signal intensity is an analog quantity of 0-5V, which is converted into a digital quantity of 0-65535 by a 16-bit ADC. The digital signal is filtered by the mean, and the average value is taken every 10 data points to obtain the first detection signal data. Similarly, the reaction signal of the anti-Sm antibody is amplified and digitally converted, and the signal range of 0-10V is mapped to the standard interval of 0-100 to obtain the second detection signal data. The detection signal of the anti-nRNP antibody is also digitized and intensity converted to form the third detection signal data. After the three sets of data are associated with the detection time information, a detection data set is formed. After quality inspection, outliers that deviate from the mean by more than three times the standard deviation are removed, and finally a unique identification code such as "20250206-000001-0001" is assigned to the qualified data to complete the generation of the original biomarker detection data set.

[0046] In a specific embodiment, if Figure 2 As shown in FIG. 1 , it is a flow chart of the steps of performing unit conversion and numerical standardization calculation on the original biomarker detection data set. The process of step S102 may specifically include the following steps:

[0047] (1) Convert the anti-dsDNA antibody detection data in the original biomarker detection data set into international units to obtain the first standard unit data;

[0048] (2) Quantifying the signal intensity of the anti-Sm antibody detection data in the original biomarker detection data set to obtain the second standard unit data;

[0049] (3) Normalizing the anti-nRNP antibody detection data in the original biomarker detection data set to obtain the third standard unit data;

[0050] (4) subtracting the baseline value from the first standard unit data and dividing it by the control value to obtain a first standardized value;

[0051] (5) Subtracting the background value from the second standard unit data and multiplying the result by the correction coefficient to obtain a second standardized value;

[0052] (6) performing linear interval interception on the third standard unit data to obtain a third standardized value;

[0053] (7) Arrange and combine the first standardized value, the second standardized value, and the third standardized value in time series to obtain a standardized biomarker data matrix.

[0054] Specifically, the anti-dsDNA antibody detection data in the original biomarker detection data set is converted according to the international unit IU / mL. The conversion process is based on the standard curve. The standard curve equation is established through the known concentration value of the reference substance, and the detection signal value is mapped to the international unit value. The standard curve adopts the multi-point regression method, selects at least 6 concentration gradient standards, and obtains the first standard unit data.

[0055] For the anti-Sm antibody detection data, signal intensity quantification involves multiple processing steps. The quantification process uses internal calibrators to correct the system error, and the original signal intensity is converted to a standard unit value by piecewise linear interpolation. In the signal quantification process, a temperature compensation factor is introduced to eliminate the influence of ambient temperature on the signal intensity, and the second standard unit data is obtained.

[0056] When normalizing the anti-nRNP antibody test data, the maximum and minimum normalization method is used. The valid range of the data is determined, obvious outliers are removed, and then the values ​​are mapped to the interval [0,1]. The systematic errors between test batches are considered in the normalization process, and the batch correction factor is used to compensate for them to obtain the third standard unit data.

[0057] The first standard unit data is standardized using the following formula:

[0058] ;

[0059] in, is the first normalized value, is the first standard unit data, is the baseline value, is the control value, is the weight coefficient of anti-dsDNA antibody, is the temperature correction factor.

[0060] The standardized calculation formula for the second standard unit data is:

[0061] ;

[0062] in, is the second normalized value, is the second standard unit data, is the background value, is the anti-Sm antibody correction factor, is the batch compensation factor, is the humidity correction factor.

[0063] The linear interval interception of the third standard unit data adopts the following formula:

[0064] ;

[0065] in, is the third normalized value, is the third standard unit data, is the lower limit of the linear interval, is the upper limit of the linear interval, is the anti-nRNP antibody correction factor, is the range calibration factor.

[0066] The three standardized values ​​are arranged in the order of detection time to form a standardized biomarker data matrix. Each row in the matrix represents the data at a time point, contains three standardized values, and each column corresponds to the detection result of an antibody. The matrix structure facilitates subsequent time series analysis and correlation calculation.

[0067] For example, when a batch of test samples is standardized, the raw data of anti-dsDNA antibodies are converted into international units through a standard curve. The standard curve is established using standards with 6 concentration points (0, 25, 50, 100, 200, 400 IU / mL). The converted first standard unit data is corrected for the baseline value and normalized for the control value to calculate the first standardized value. Similarly, the detection data of anti-Sm antibodies is quantified for signal intensity and corrected for background value, and then multiplied by the correction coefficient to obtain the second standardized value. The anti-nRNP antibody data is normalized and linear interval intercepted to ensure that the value distribution is within the valid range. Finally, the three standardized values ​​are organized into a matrix form according to the detection time to complete the data standardization.

[0068] In a specific embodiment, if Figure 3 As shown in FIG. 1 , it is a timing diagram of the steps of calculating the correlation coefficient and the numerical distribution characteristics between the detection values ​​of the three antibodies. The process of executing step S103 may specifically include the following steps:

[0069] (1) Dividing the standardized biomarker data matrix into data subsets according to the time window, calculating the correlation coefficient between anti-dsDNA antibodies and anti-Sm antibodies for each data subset, and obtaining the first correlation coefficient sequence;

[0070] (2) performing sliding window calculation on the detection values ​​of anti-Sm antibody and anti-nRNP antibody in the standardized biomarker data matrix to obtain a second correlation coefficient sequence;

[0071] (3) Calculating the detection values ​​of anti-dsDNA antibodies and anti-nRNP antibodies in the standardized biomarker data matrix in segments to obtain a third correlation coefficient sequence;

[0072] (4) combining the first correlation coefficient sequence, the second correlation coefficient sequence, and the third correlation coefficient sequence to form a correlation matrix;

[0073] (5) Performing mean calculation and variance analysis on the first correlation coefficient sequence in the correlation array to obtain the fluctuation range data of anti-dsDNA antibody and anti-Sm antibody;

[0074] (6) Calculate the quartiles of the second correlation coefficient sequence in the correlation array to obtain the distribution interval data of anti-Sm antibodies and anti-nRNP antibodies;

[0075] (7) Calculating the kurtosis coefficient and skewness coefficient of the third correlation coefficient sequence in the correlation array to obtain the distribution morphology data of anti-dsDNA antibodies and anti-nRNP antibodies;

[0076] (8) Combining the fluctuation range data, distribution interval data, and distribution form data to obtain distribution characteristic data;

[0077] (9) Combine the correlation array with the distribution feature data to obtain the biomarker association feature vector.

[0078] Specifically, the standardized biomarker data matrix was divided into time windows. The time window size was set to 14 days, and each time it slid back 7 days to form a series of overlapping data subsets. The Pearson correlation coefficient was calculated for the anti-dsDNA antibody and anti-Sm antibody detection values ​​in each data subset. The correlation coefficient was calculated using the following formula:

[0079] ;

[0080] in, is the correlation coefficient between anti-dsDNA antibody and anti-Sm antibody, and are the normalized values ​​of the two antibodies at time point i, and is the respective mean, and n is the number of data points. The correlation coefficients of each time window are arranged in chronological order to obtain the first correlation coefficient sequence.

[0081] The detection values ​​of anti-Sm antibody and anti-nRNP antibody were analyzed by sliding window, with a window size of 10 days and a sliding step of 5 days. The correlation coefficient calculation formula in each window is:

[0082] ;

[0083] in, is the correlation coefficient of the two antibodies, and is the detection value within the window, and is the mean value within the window, is the window weight factor, is the sliding compensation coefficient, and m is the number of data points in the window.

[0084] The detection values ​​of anti-dsDNA antibodies and anti-nRNP antibodies were calculated by segmentation, with each segment lasting 20 days. The segmentation correlation coefficient was calculated by weighted method:

[0085] ;

[0086] in, is the segment correlation coefficient, is the weight coefficient of the kth segment, and is the detection value within the segment, is the segment correction factor, is the time attenuation coefficient, s is the total number of segments, l is the number of data points in each segment.

[0087] After combining the three correlation coefficient sequences into a correlation array, calculate the mean and variance of the first correlation coefficient sequence:

[0088] ;

[0089] in, is the mean, is the variance, is the correlation coefficient value in the sequence, is the mean correction factor, is the variance correction coefficient, and N is the sequence length.

[0090] The quartile calculation formula for the second correlation coefficient sequence is:

[0091] ;

[0092] in, is the p quantile value, n is the sample size, is the quantile adjustment factor, It represents the ith observation value after sorting the correlation coefficient sequence from small to large, and p represents the probability value of the quantile (such as 0.25, 0.5, 0.75).

[0093] The calculation formula for the kurtosis and skewness of the third correlation coefficient sequence is:

[0094] ;

[0095] in, The kurtosis coefficient represents the correlation coefficient sequence between anti-dsDNA antibody and anti-nRNP antibody; represents the skewness coefficient of the correlation coefficient sequence between anti-dsDNA antibody and anti-nRNP antibody; n represents the number of samples in the correlation coefficient sequence; represents the i-th correlation coefficient value; Represents the arithmetic mean of the correlation coefficient sequence; represents the kurtosis adjustment factor, which is used to correct the peak degree of the distribution; Represents the skewness adjustment factor, which is used to correct the asymmetry of the distribution.

[0096] For example: a set of 90 consecutive days of test data is analyzed, and the data subsets are divided according to the 14-day window size, sliding once every 7 days to obtain 12 overlapping data subsets. The correlation coefficients of anti-dsDNA antibodies and anti-Sm antibodies are calculated in each subset to form the first correlation coefficient sequence. Similarly, a 10-day window is used to perform sliding analysis on the anti-Sm antibody and anti-nRNP antibody data to obtain the second correlation coefficient sequence. The 90-day data is divided into 4 segments of 20 days, and the correlation coefficients of anti-dsDNA antibodies and anti-nRNP antibodies are calculated respectively to form the third correlation coefficient sequence. After the three sequences are combined into a correlation matrix, statistical feature calculations are performed separately, including mean, variance, quartiles, kurtosis and skewness coefficients. Finally, all feature data are integrated into a biomarker-associated feature vector to complete multi-dimensional correlation analysis.

[0097] In a specific embodiment, the process of executing step S104 may specifically include the following steps:

[0098] (1) Split the biomarker-associated feature vector into multiple time period data according to the time series, calculate the change rate of each time period data, and obtain a dynamic change rate sequence;

[0099] (2) Detect extreme points and divide the intervals of the dynamic change rate sequence to obtain the detection value fluctuation interval data;

[0100] (3) Mark the rising interval and the falling interval in the detection value fluctuation interval data respectively, calculate the slope value of each interval, and obtain the trend change data;

[0101] (4) Perform periodic analysis on trend change data, count the duration of rising and falling intervals, and obtain periodic characteristic data;

[0102] (5) Correlate the change amplitude and change duration in the periodic characteristic data to obtain the change intensity data;

[0103] (6) Integrate the dynamic change rate series, detection value fluctuation range data, trend change data, periodic characteristic data, and change intensity data to obtain a biomarker detection evaluation report.

[0104] Specifically, the biomarker-associated feature vectors were split into time periods of 30 days, with adjacent time periods overlapping for 15 days to form a continuous data sequence. The rate of change was calculated for the data in each time period. The rate of change was calculated using the central difference method, that is, the value at the current moment was subtracted from the value at the previous moment, and then divided by the time interval to obtain a dynamic rate of change sequence.

[0105] The extreme point detection of the dynamic change rate sequence adopts the three-point method, that is, the current point is compared with the two points before and after it. If the current point is greater than the two points before and after, it is marked as a maximum point; if the current point is less than the two points before and after, it is marked as a minimum point. The entire sequence is divided into several fluctuation intervals by the identified extreme points. Each interval is defined by adjacent maximum and minimum points to form the detection value fluctuation interval data. For the processing of fluctuation interval data, it is necessary to identify the rising interval and the falling interval respectively. The rising interval refers to the interval from the minimum point to the maximum point, and the falling interval refers to the interval from the maximum point to the minimum point. The slope value is calculated for each interval, and the slope value is obtained by dividing the numerical difference between the two end points of the interval by the time interval. The slope of the rising interval is positive, and the slope of the falling interval is negative. These slope values ​​constitute trend change data.

[0106] Periodic analysis is an important step in conducting in-depth research on trend change data. First, count the duration of the rising and falling intervals, that is, the time span from the start to the end of the interval. By recording the start and end time points of each interval, calculate the duration of the interval. At the same time, analyze the alternation of adjacent intervals, record the rising-falling cycle, and obtain periodic characteristic data.

[0107] The calculation of change intensity data requires comprehensive consideration of two factors: the amplitude of change and the duration. The amplitude of change refers to the amount of change in the value within the interval, that is, the absolute value of the interval end value minus the starting value. The amplitude of change and the duration are weighted and calculated, and the weight coefficient is determined according to clinical practice. A larger amplitude of change and a shorter duration correspond to a higher intensity of change. All analyzed data are integrated. The dynamic change rate sequence reflects the instantaneous change characteristics of the indicator, the detection value fluctuation interval data shows the fluctuation range of the value, the trend change data characterizes the direction and rate of change, the periodic characteristic data reveals the regularity of the change, and the change intensity data quantifies the severity of the change. These data are organized in a unified format to form a biomarker detection evaluation report.

[0108] For example, the test data of a SLE patient for 180 consecutive days is analyzed. First, the data is split into 30-day time periods to obtain 6 time periods, and each time period overlaps with the adjacent period by 15 days. The dynamic change rate of each time point is calculated to form a change rate sequence. Through extreme point detection, the peaks and troughs in the sequence are identified, and the entire sequence is divided into multiple fluctuation intervals. Each fluctuation interval is analyzed, the rising interval and the falling interval are marked, and the slope value of each interval is calculated. Statistical analysis shows that the average duration of the rising interval is 12 days, and the average duration of the falling interval is 15 days, which constitutes a complete cycle of about 27 days. Combined with the analysis of the change amplitude and duration, the change intensity of each interval is calculated. All analysis data are integrated to generate an evaluation report containing all characteristic indicators to provide data support for clinical diagnosis and treatment.

[0109] In a specific embodiment, the process of executing step S105 may specifically include the following steps:

[0110] (1) Divide the test values ​​in the biomarker test evaluation report into time series, calculate the mean and standard deviation of the values ​​in each time period, and obtain the benchmark statistical data;

[0111] (2) Perform outlier detection on the values ​​in the benchmark statistical data, mark the data points that are beyond three times the standard deviation, and obtain the outlier location data;

[0112] (3) Cluster the adjacent outliers in the outlier location data, calculate the duration of the outlier interval, and obtain the outlier duration data;

[0113] (4) Quantify the amplitude of the outliers in the abnormal continuous data, calculate the degree to which the outliers deviate from the mean, and obtain the abnormal degree data;

[0114] (5) Perform correlation analysis on the abnormal point location data, abnormal duration data, and abnormal degree data, calculate the frequency of abnormal features, and obtain abnormal pattern data;

[0115] (6) Arrange the abnormal pattern data in chronological order, integrate the location, duration, and degree information of the abnormal features, and obtain the detection data warning information.

[0116] Specifically, when performing abnormal analysis on the data in the biomarker test assessment report, time series segmentation is required. The test data is divided into time periods of 30 days, and the mean and standard deviation of the data in each time period are calculated. The mean is calculated using the arithmetic mean method, which sums all the test values ​​in the time period and divides it by the number of data points; the standard deviation is calculated by calculating the sum of the squares of the deviations of each data point from the mean, dividing it by the number of data points minus 1 and then taking the square root, thereby forming benchmark statistical data.

[0117] The 3σ criterion is used for outlier detection of benchmark statistical data. The deviation of each data point from the mean is calculated, and the deviation value is divided by the standard deviation. If the quotient is greater than 3 or less than -3, the data point is marked as an outlier. This method is based on the properties of normal distribution and marks outliers that deviate significantly from the normal range. After marking, the time position and value of each outlier are recorded to obtain the outlier location data. The cluster analysis of outliers adopts a distance-based method. The time interval threshold is set to 3 days. If the time interval between two outliers is less than the threshold, the two outliers are considered to belong to the same outlier interval. All outliers are scanned in turn, and adjacent outliers are combined into outlier intervals. The start time and end time of each outlier interval are calculated to obtain the outlier duration data.

[0118] When quantifying the amplitude of outliers, it is necessary to calculate the degree of deviation between the outliers and the mean. The specific method is to subtract the mean of the corresponding time period from the outlier, and then divide it by the standard deviation to obtain the standardized deviation value. For continuous abnormal intervals, calculate the average deviation of all abnormal points in the interval as the degree of abnormality of the abnormal interval, so as to obtain the degree of abnormality data. The correlation analysis of abnormal features needs to comprehensively consider the three dimensions of location, duration and degree. First, count the distribution frequency of abnormal points in different time periods, and record the duration and average degree of abnormality of each abnormal interval. Then analyze the correlation between abnormal features, including the time interval between adjacent abnormal intervals, the trend of changes in the degree of abnormality, etc., to obtain abnormal pattern data.

[0119] The last step is the time series integration of abnormal pattern data, which is particularly important. For each detected abnormal interval, record its specific start and end time points, the number of days it lasts, and the average deviation of the abnormal values ​​in the interval. At the same time, mark the type of abnormality, such as single-point abnormality or continuous abnormality, rising abnormality or falling abnormality. Arrange this information in chronological order and construct a data table containing complete abnormal characteristics. This table contains key information such as the time location of the abnormality, the duration span, and the quantitative value of the abnormality degree, forming early warning information for detection data.

[0120] For example, the 180-day continuous monitoring data of a SLE patient was analyzed. The data was divided into 6 time periods of 30 days, and the mean and standard deviation of each time period were calculated as the baseline value. The abnormal points were identified by the 3σ criterion. For example, continuous abnormal high values ​​were detected on the 45th, 47th, and 48th days, and these three points were clustered into an abnormal interval. The average abnormal degree (standard deviation multiple) of the abnormal interval was calculated and recorded as 3.5 times the standard deviation; at the same time, the duration of the abnormal interval was recorded as 4 days. Further analysis found that there was a single point abnormality on the 120th and 150th days, with abnormal degrees of 3.2 times and 3.8 times the standard deviation, respectively. This information was organized into an early warning information table, including the start and end time of the abnormal interval (45-48 days), duration (4 days), average abnormal degree (3.5 times the standard deviation), and complete information of two single point abnormalities.

[0121] In a specific embodiment, the process of executing step S106 may specifically include the following steps:

[0122] (1) Divide the detection data warning information into multiple observation windows according to time periods, calculate the warning frequency in each window, and obtain the warning frequency data;

[0123] (2) Perform time series analysis on the warning frequency data, calculate the change in warning frequency between adjacent time windows, and obtain warning trend data;

[0124] (3) Mark the upward trend and downward trend in the warning trend data separately, calculate the duration interval of the trend change, and obtain the trend continuity data;

[0125] (4) Conduct periodic analysis on trend persistence data, statistically analyze the occurrence patterns of different types of trends, and obtain trend pattern data;

[0126] (5) Conduct comprehensive correlation analysis on warning frequency data, warning trend data, trend duration data, and trend regularity data to obtain dynamic feature data;

[0127] (6) Integrate and analyze the various characteristic indicators in the dynamic characteristic data to form the biomarker detection data analysis results.

[0128] Specifically, the analysis of the warning information of the detection data first requires time window division. The warning information is divided into 15 days as an observation window, and adjacent windows overlap for 5 days to ensure the continuity of the data. In each observation window, the number of warning events is counted and the warning frequency is calculated. The calculation method of the warning frequency is to divide the total number of warning events in the window by the time length of the window to obtain the average number of warnings per day to form the warning frequency data. The time series analysis of the warning frequency data is achieved by calculating the change in warning frequency between adjacent observation windows. For each pair of adjacent windows, the warning frequency of the latter window is subtracted from the warning frequency of the previous window to obtain the change value of the warning frequency. A positive change value indicates an increase in the warning frequency, and a negative change value indicates a decrease in the warning frequency. The change values ​​of all windows are arranged in chronological order to form warning trend data.

[0129] In the processing of warning trend data, the upward trend and downward trend are marked separately. Continuous positive change values ​​constitute an upward trend interval, and continuous negative change values ​​constitute a downward trend interval. For each trend interval, calculate its duration, that is, the number of days from the beginning to the end of the trend. At the same time, record the cumulative change in the warning frequency in each interval to obtain the trend continuity data. The periodic analysis of trend continuity data includes statistics on the occurrence patterns of different types of trends. Analyze the alternation of upward and downward trends, and record the duration of a complete cycle (one rise and one fall constitute a cycle). Count the occurrence frequency of trend intervals of different durations, analyze the time characteristics of trend changes, and obtain trend regularity data.

[0130] In the comprehensive correlation analysis stage, all the acquired data need to be analyzed in multiple dimensions. The warning frequency data is matched with the trend data, and the relationship between the high-frequency warning period and the trend change is analyzed; the trend duration data is compared with the regular data to identify the typical change pattern. Through these analyses, dynamic feature data reflecting the overall change characteristics are obtained. In the final integrated analysis stage, the various indicators in the dynamic feature data are systematically organized. Including the basic statistical characteristics of the warning frequency (average value, fluctuation range), the typical characteristics of the trend change (the duration distribution of the rising / falling trend), and the periodic characteristics (the main cycle length and its stability), to form the analysis results of the biomarker detection data.

[0131] For example, a SLE patient was continuously monitored for 90 days, and the warning information of the detection data was divided into windows of 15 days, resulting in 10 overlapping observation windows. The number of warning events was counted in each window, such as 3 warnings in the first window, 5 warnings in the second window, and so on, to calculate the warning frequency of each window. The warning trend change was obtained by calculating the difference in warning frequencies of adjacent windows. The analysis found that the warning frequency of 3 consecutive windows increased from the 30th day, followed by a downward trend of 2 windows, forming a change cycle. This rise-fall cycle occurred a total of 3 times in 90 days, indicating that the warning frequency has a periodic change characteristic of about 30 days. In the rising trend interval, the cumulative increase of the warning frequency is more significant, while the change in the falling trend interval is relatively gentle.

[0132] In a specific embodiment, the process of performing the step of marking the rising trend and the falling trend in the early warning trend data respectively may specifically include the following steps:

[0133] (1) The warning trend data is segmented according to the direction of value change, marked as rising segments and falling segments, and trend segmented data is obtained;

[0134] (2) Calculate the slope value for each segment in the trend segmented data, divide the change rate level according to the slope size, and obtain the trend rate data;

[0135] (3) Compare the change rates of adjacent segments in the trend rate data, mark the acceleration and deceleration intervals of the change rate, and obtain the rate change data;

[0136] (4) Arrange the change intervals in the rate change data in chronological order, count the duration of each interval, and obtain the interval duration data;

[0137] (5) Perform distribution statistics on the interval duration data, calculate the interval distribution characteristics of different change types, and obtain the interval distribution data;

[0138] (6) Combine the trend segmentation data, trend rate data, rate change data, interval duration data, and interval distribution data to obtain trend continuity data.

[0139] Specifically, by calculating the numerical difference between adjacent time points, the direction of change at each time point is determined. When the direction of change of multiple consecutive time points is the same, these points are classified as the same segment. Specifically, if the difference between two adjacent points is greater than zero, it is marked as an ascending segment; if the difference is less than zero, it is marked as a descending segment; if the difference is equal to zero, it is consistent with the previous segment. The trend segmented data obtained in this way contains the information of ascending and descending segments. When calculating the slope value for each segment of the trend segmented data, the method of the first and last points in the segment is used. The value of the end point of the segment is subtracted from the value of the starting point, and then divided by the corresponding time interval to obtain the average rate of change of the segment. According to the absolute value of the slope, the rate of change is divided into different levels. The rate level division standard is set based on clinical experience, such as defining the absolute value of the slope between 0-0.5 as slow change, 0.5-1.0 as medium-speed change, and greater than 1.0 as rapid change, thereby obtaining trend rate data.

[0140] The comparative analysis of adjacent segments of trend rate data focuses on the progressive relationship of the rate of change. Calculate the difference in the slope values ​​of two adjacent segments. If the absolute value of the slope of the latter segment is greater than that of the previous segment, it is marked as an acceleration interval; if it is less than that of the previous segment, it is marked as a deceleration interval. This comparison takes into account both acceleration and deceleration in the same direction (such as acceleration or deceleration in an upward trend) and rate changes during turning to obtain rate change data. When arranging the rate change data in time series, it is necessary to retain the complete information of each change interval. Record the starting and ending time points of each interval and calculate the duration of the interval. At the same time, mark the change type (acceleration or deceleration) and change rate level of the interval. This information constitutes the interval duration data.

[0141] The distribution statistics of interval duration data need to be classified and summarized. Count the number distribution of acceleration intervals and deceleration intervals separately, and calculate the frequency of occurrence of different durations. At the same time, analyze the distribution characteristics of intervals of different rate levels, including the average duration and frequency of occurrence of each type of interval, to obtain interval distribution data. Organize all the data obtained previously in a systematic way. Use trend segmentation data as the basic framework, and annotate the rate level, rate change characteristics (acceleration or deceleration), duration, and position characteristics of each segment in the overall distribution. This multi-dimensional data combination constitutes trend duration data.

[0142] For example: the monitoring data of the anti-dsDNA antibody level of a SLE patient was analyzed. First, the direction of the numerical change was identified, and it was found that there was a continuous upward trend from the 7th day to the 14th day from the start of monitoring, and then a downward trend from the 15th day to the 20th day. The slope value of the rising segment was calculated, and it was determined to be a rapid rising interval based on the slope size; the slope value of the falling segment was also calculated and determined to be a medium-speed falling interval. By comparing the absolute values ​​of the slopes of these two adjacent segments, it can be determined that the process from rising to falling is a deceleration change. The duration of these two segments was recorded as 8 days and 6 days respectively. Through statistical analysis of all similar segments, it was found that during the entire monitoring period, the average duration of the rapid change interval was shorter, while the duration of the medium-speed and slow-speed change intervals was longer. This distribution feature reflects the typical law of changes in biomarker levels.

[0143] In a specific embodiment, the process of performing the step of periodically analyzing the trend persistence data may specifically include the following steps:

[0144] (1) Divide the trend persistence data into segments according to fixed time intervals, mark the distribution positions of the upward and downward trends in each segment, and obtain the trend distribution data;

[0145] (2) Calculate the time intervals of trend change points in the trend distribution data, count the time differences between the change points, and obtain time series data;

[0146] (3) Group the time series data according to the law of change, calculate the mean and variance of the time interval of each group, and obtain the periodic characteristic data;

[0147] (4) Perform frequency statistics on the time intervals in the periodic characteristic data, calculate the number of occurrences of different period lengths, and obtain periodic frequency data;

[0148] (5) Sort the periodic frequency data according to the frequency of occurrence, filter out the main periodic patterns and the secondary periodic patterns, and obtain the periodic pattern data;

[0149] (6) Integrate trend distribution data, time series data, periodic characteristic data, periodic frequency data, and periodic pattern data to obtain trend regularity data.

[0150] Specifically, the data is divided into fixed time periods of 30 days, and the starting and ending positions of the upward and downward trends are marked in each time period. The intervals where the upward trend index value continuously increases and the intervals where the downward trend index value continuously decreases are recorded to form trend distribution data. The trend change points in the trend distribution data reflect the turning point of the value change direction. The specific time of each change point is recorded, and the time interval between adjacent change points is calculated. The change points include the peak point from rising to falling and the valley point from falling to rising. The time difference between these points is recorded to obtain a series of time series data reflecting the change cycle.

[0151] The grouping analysis of time series data is based on the similarity of the changing patterns. Similar time intervals are grouped together, for example, intervals of 15-20 days are grouped as short-cycle groups, intervals of 21-30 days are grouped as medium-cycle groups, and intervals of more than 30 days are grouped as long-cycle groups. The mean is calculated for each group of data to reflect the typical length of the cycle in that group; the variance is calculated to reflect the stability of the cycle length. These statistical characteristics constitute the cycle characteristic data. The frequency statistics of cycle characteristic data require calculating the number of occurrences of each cycle length. All observed cycles are grouped according to length, and the frequency of occurrence of each group of cycles during the entire monitoring period is counted. The statistical results include the number of occurrences of various cycle lengths and their proportion in the total number of cycles, forming cycle frequency data.

[0152] The sorting analysis of cycle frequency data focuses on identifying the dominant cycle. Different cycles are sorted from high to low according to the frequency of occurrence. The cycle with the highest frequency is defined as the primary cycle pattern, and the second highest is defined as the secondary cycle pattern. This classification reflects the main rhythmic characteristics of the changes in biomarker levels and is recorded as cycle pattern data. The last step is the integration of all analysis data. Using the trend distribution data as the basic framework, mark the cycle type to which each trend interval belongs, record the characteristic parameters of the cycle (such as mean, variance), and indicate its position in the overall cycle distribution (primary mode or secondary mode). This multi-level data integration forms trend regularity data.

[0153] For example, a SLE patient was continuously monitored for 120 days. The data was first divided into four time periods of 30 days. Three upward trends and two downward trends were identified in the first 30 days, and the start and end time of each trend was recorded. The time intervals between adjacent trend change points were calculated, and it was found that the duration of the first complete cycle (one rise plus one decline) was 12 days, the second cycle was 15 days, and the third cycle was 14 days. These time interval data were grouped and statistically analyzed, and it was found that the 12-15 day cycle was the most common, constituting the main cycle pattern; at the same time, there were a small number of longer cycles of 20-25 days, constituting the secondary cycle pattern. This cycle analysis reveals the regular characteristics of changes in biomarker levels and provides an important reference for subsequent monitoring. During the entire monitoring process, the frequency of the main cycle pattern was significantly higher than that of the secondary cycle pattern, indicating that the patient's biomarker level has a relatively stable change pattern.

[0154] The above describes the biomarker detection method for systemic lupus erythematosus in the embodiments of the present application. The following describes the biomarker detection system for systemic lupus erythematosus in the embodiments of the present application. Figure 4 In the embodiments of the present application, an embodiment of a biomarker detection system for systemic lupus erythematosus includes:

[0155] The acquisition module 201 is used to connect the biomarker detection device through the detection data acquisition module, collect the detection data of anti-dsDNA antibody, anti-Sm antibody and anti-nRNP antibody in real time, and obtain the original detection data set of the biomarker;

[0156] A conversion module 202 is used to perform unit uniform conversion and numerical standardization calculation on the original biomarker detection data set to obtain a standardized biomarker data matrix;

[0157] A calculation module 203 is used to calculate the correlation coefficient and numerical distribution characteristics between the three antibody detection values ​​based on the standardized biomarker data matrix using a multidimensional data association analysis algorithm to obtain a biomarker association feature vector;

[0158] An evaluation module 204 is used to analyze the dynamic change characteristics of the detection value according to the biomarker-related feature vector to obtain a biomarker detection evaluation report;

[0159] The identification module 205 is used to perform feature pattern recognition on abnormal changes in the detection value according to the biomarker detection evaluation report to obtain detection data warning information;

[0160] The analysis module 206 is used to perform dynamic feature analysis on the biomarker detection data based on the detection data warning information to obtain a biomarker detection data analysis result.

[0161] Through the collaboration of the above components, the real-time data collection of the biomarker detection equipment is realized, ensuring the timely acquisition and integrity of the detection data of the three biomarkers, anti-dsDNA antibody, anti-Sm antibody and anti-nRNP antibody. By uniformly converting the units and calculating the numerical standardization of the original biomarker detection data set, the comparability of data under different batches and different detection conditions is solved, and the standardized biomarker data matrix is ​​obtained, which lays the foundation for subsequent analysis. The multidimensional data association analysis algorithm is used to calculate the correlation coefficient and numerical distribution characteristics between the detection values ​​of the three antibodies, which not only realizes the joint analysis of multiple biomarker data, but also automatically extracts the association characteristics between the data through the algorithm. The obtained biomarker association feature vector can fully reflect the distribution law of the data. The dynamic change characteristics of the detection values ​​are analyzed, and the biomarker detection evaluation report is generated, realizing the quantitative description of the data change trend. The abnormal changes of the detection values ​​are analyzed by feature pattern recognition technology, and a data-based early warning mechanism is established. The obtained detection data early warning information is objective and reliable. Finally, through dynamic feature analysis, the biomarker detection data was comprehensively analyzed, and the analysis results obtained covered multiple dimensional characteristics of the data, which not only improved the automation level of data processing, but also enhanced the depth and accuracy of data analysis.

[0162] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described systems, systems and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0163] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for detecting a biomarker for systemic lupus erythematosus, characterized in that: The biomarker detection method for systemic lupus erythematosus comprises: The biomarker detection equipment is connected via the detection data acquisition module to collect the detection data of anti-dsDNA antibodies, anti-Sm antibodies and anti-nRNP antibodies in real time to obtain the original detection data set of the biomarker; Performing unit uniform conversion and numerical standardization calculation on the original biomarker detection data set to obtain a standardized biomarker data matrix; Based on the standardized biomarker data matrix, a multidimensional data association analysis algorithm is used to calculate the correlation coefficients and numerical distribution characteristics between the detection values ​​of the three antibodies to obtain a biomarker association feature vector, including: dividing the standardized biomarker data matrix into data subsets according to a time window, calculating the correlation coefficients of anti-dsDNA antibodies and anti-Sm antibodies for each data subset, and obtaining a first correlation coefficient sequence; performing a sliding window calculation on the detection values ​​of anti-Sm antibodies and anti-nRNP antibodies in the standardized biomarker data matrix to obtain a second correlation coefficient sequence; performing segmented calculation on the detection values ​​of anti-dsDNA antibodies and anti-nRNP antibodies in the standardized biomarker data matrix to obtain a third correlation coefficient sequence; and combining the first correlation coefficient sequence, the first correlation coefficient sequence, the second correlation coefficient sequence, and the third correlation coefficient sequence. The second correlation coefficient sequence and the third correlation coefficient sequence are combined to form a correlation array; the first correlation coefficient sequence in the correlation array is subjected to mean calculation and variance analysis to obtain the fluctuation range data of anti-dsDNA antibodies and anti-Sm antibodies; the second correlation coefficient sequence in the correlation array is subjected to quartile calculation to obtain the distribution interval data of anti-Sm antibodies and anti-nRNP antibodies; the kurtosis coefficient and skewness coefficient of the third correlation coefficient sequence in the correlation array are calculated to obtain the distribution morphology data of anti-dsDNA antibodies and anti-nRNP antibodies; the fluctuation range data, distribution interval data and distribution morphology data are combined to obtain distribution characteristic data; the correlation array is combined with the distribution characteristic data to obtain a biomarker association feature vector; Analyzing the dynamic change characteristics of the detection value according to the biomarker-associated feature vector to obtain a biomarker detection evaluation report; According to the biomarker detection evaluation report, characteristic pattern recognition is performed on abnormal changes in detection values ​​to obtain detection data warning information; Based on the detection data warning information, dynamic feature analysis is performed on the biomarker detection data to obtain a biomarker detection data analysis result.

2. The biomarker detection method for systemic lupus erythematosus according to claim 1, characterized in that: The detection data acquisition module is connected to the biomarker detection device to collect the detection data of anti-dsDNA antibodies, anti-Sm antibodies and anti-nRNP antibodies in real time to obtain the original biomarker detection data set, including: Receiving the fluorescence intensity signal of the anti-dsDNA antibody through the biomarker detection device, digitally converting the fluorescence signal, and generating first detection signal data; receiving a reaction signal of the anti-Sm antibody through a biomarker detection device, performing intensity conversion on the reaction signal, and generating second detection signal data; receiving a reaction signal of the anti-nRNP antibody through a biomarker detection device, performing intensity conversion on the reaction signal, and generating third detection signal data; Associating and marking the first detection signal data, the second detection signal data, the third detection signal data and the detection time information to generate a detection data group; The test data set is quality checked to remove abnormal data points to obtain qualified test data; the qualified test data is assigned a unique identification code to generate the biomarker original test data set.

3. The biomarker detection method for systemic lupus erythematosus according to claim 1, characterized in that: The unit conversion and numerical standardization calculation of the original biomarker detection data set are performed to obtain a standardized biomarker data matrix, including: Converting the anti-dsDNA antibody detection data in the original biomarker detection data set according to international units to obtain first standard unit data; quantifying the signal intensity of the anti-Sm antibody detection data in the original detection data set of the biomarker to obtain second standard unit data; Normalizing the anti-nRNP antibody detection data in the original biomarker detection data set to obtain third standard unit data; Subtract the baseline value from the first standard unit data and divide the result by the control value to obtain a first standardized value; Subtract the background value from the second standard unit data and multiply by the correction coefficient to obtain a second standardized value; Performing linear interval interception on the third standard unit data to obtain a third standardized value; The first standardized value, the second standardized value, and the third standardized value are arranged and combined in time series to obtain a standardized biomarker data matrix.

4. The biomarker detection method for systemic lupus erythematosus according to claim 1, characterized in that: The biomarker detection evaluation report is obtained by analyzing the dynamic change characteristics of the detection value based on the biomarker associated feature vector, including: Splitting the biomarker-associated feature vector into multiple time period data according to the time series, calculating the change rate of each time period data, and obtaining a dynamic change rate sequence; Perform extreme point detection and interval division on the dynamic change rate sequence to obtain detection value fluctuation interval data; Mark the rising interval and the falling interval in the detection value fluctuation interval data respectively, calculate the slope value of each interval, and obtain the trend change data; Perform periodic analysis on the trend change data, count the duration of the rising interval and the falling interval, and obtain periodic characteristic data; Correlate and calculate the change amplitude and change duration in the periodic characteristic data to obtain change intensity data; The dynamic change rate sequence, detection value fluctuation interval data, trend change data, periodic characteristic data, and change intensity data are integrated to obtain a biomarker detection evaluation report.

5. The biomarker detection method for systemic lupus erythematosus according to claim 1, characterized in that: According to the biomarker detection evaluation report, feature pattern recognition is performed on abnormal changes in detection values ​​to obtain detection data warning information, including: The detection values ​​in the biomarker detection evaluation report are segmented according to the time series, and the mean and standard deviation of the values ​​in each time period are calculated to obtain the benchmark statistical data; Performing outlier detection on the values ​​in the benchmark statistical data, marking data points that exceed three times the standard deviation, and obtaining outlier point location data; Clustering adjacent abnormal points in the abnormal point location data, calculating the duration of the abnormal interval, and obtaining abnormal duration data; quantifying the amplitude of the abnormal values ​​in the abnormal continuous data, calculating the degree to which the abnormal values ​​deviate from the mean, and obtaining abnormal degree data; Perform correlation analysis on the abnormal point location data, abnormal duration data, and abnormal degree data, and calculate the frequency of occurrence of abnormal features to obtain abnormal pattern data; The abnormal pattern data are arranged in chronological order, and the location, duration, and degree information of the abnormal features are integrated to obtain detection data warning information.

6. The biomarker detection method for systemic lupus erythematosus according to claim 1, characterized in that: The method of performing dynamic feature analysis on the biomarker detection data based on the detection data warning information to obtain the biomarker detection data analysis results includes: Divide the detection data warning information into multiple observation windows according to time periods, calculate the warning frequency in each window, and obtain warning frequency data; Performing time series analysis on the warning frequency data, calculating the warning frequency changes between adjacent time windows, and obtaining warning trend data; Mark the rising trend and the falling trend in the early warning trend data respectively, calculate the duration interval of the trend change, and obtain the trend duration data; Perform periodic analysis on the trend persistence data, count the occurrence patterns of different types of trends, and obtain trend pattern data; perform comprehensive correlation analysis on the warning frequency data, warning trend data, trend persistence data, and trend pattern data to obtain dynamic feature data; The various characteristic indicators in the dynamic characteristic data are integrated and analyzed to form a biomarker detection data analysis result.

7. The method for detecting biomarkers for systemic lupus erythematosus according to claim 6, characterized in that: The step of marking the rising trend and the falling trend in the early warning trend data respectively, calculating the duration interval of the trend change, and obtaining the trend duration data includes: The warning trend data is segmented according to the direction of value change, and marked as rising segments and falling segments respectively, to obtain trend segmented data; Calculating the slope value for each segment in the trend segmented data, dividing the change rate level according to the slope size, and obtaining trend rate data; Comparing the change rates of adjacent sections in the trend rate data, marking the acceleration and deceleration intervals of the change rate, and obtaining rate change data; Arrange the change intervals in the rate change data in chronological order, count the duration of each interval, and obtain interval duration data; Performing distribution statistics on the interval duration data, calculating interval distribution characteristics of different change types, and obtaining interval distribution data; The trend segmentation data, trend rate data, rate change data, interval duration data, and interval distribution data are combined to obtain the trend continuity data.

8. The method for detecting biomarkers for systemic lupus erythematosus according to claim 6, characterized in that: The trend continuous data is periodically analyzed to count the occurrence patterns of different types of trends to obtain trend pattern data, including: The trend persistence data is segmented according to fixed time intervals, and the distribution positions of the rising trend and the falling trend in each segment are marked to obtain trend distribution data; Calculating the time intervals of trend change points in the trend distribution data, and counting the time differences between the change points to obtain time series data; The time series data are grouped according to the change rules, and the mean and variance of the time intervals of each group are calculated to obtain periodic characteristic data; Performing frequency statistics on the time intervals in the periodic characteristic data, calculating the number of occurrences of different period lengths, and obtaining periodic frequency data; Sorting the periodic frequency data according to the frequency of occurrence, screening the main periodic patterns and the secondary periodic patterns, and obtaining periodic pattern data; The trend distribution data, time series data, periodic characteristic data, periodic frequency data, and periodic pattern data are integrated to obtain the trend regularity data.

9. A biomarker detection system for systemic lupus erythematosus, used to implement the biomarker detection method for systemic lupus erythematosus according to any one of claims 1 to 8, characterized in that: The biomarker detection system for systemic lupus erythematosus comprises: The acquisition module is used to connect the biomarker detection device through the detection data acquisition module, collect the detection data of anti-dsDNA antibodies, anti-Sm antibodies and anti-nRNP antibodies in real time, and obtain the original detection data set of the biomarker; A conversion module, used to perform unit uniform conversion and numerical standardization calculation on the original biomarker detection data set to obtain a standardized biomarker data matrix; A calculation module is used to calculate the correlation coefficients and numerical distribution characteristics between the detection values ​​of three antibodies based on the standardized biomarker data matrix using a multidimensional data association analysis algorithm to obtain a biomarker association feature vector, including: dividing the standardized biomarker data matrix into data subsets according to a time window, calculating the correlation coefficients of anti-dsDNA antibodies and anti-Sm antibodies for each data subset, and obtaining a first correlation coefficient sequence; performing a sliding window calculation on the detection values ​​of anti-Sm antibodies and anti-nRNP antibodies in the standardized biomarker data matrix to obtain a second correlation coefficient sequence; performing a segmented calculation on the detection values ​​of anti-dsDNA antibodies and anti-nRNP antibodies in the standardized biomarker data matrix to obtain a third correlation coefficient sequence; performing a segmented calculation on the detection values ​​of anti-dsDNA antibodies and anti-nRNP antibodies in the standardized biomarker data matrix to obtain a third correlation coefficient sequence; performing a segmented calculation on the first correlation coefficient sequence. The first correlation coefficient sequence, the second correlation coefficient sequence, and the third correlation coefficient sequence are combined to form a correlation array; the first correlation coefficient sequence in the correlation array is subjected to mean calculation and variance analysis to obtain the fluctuation range data of anti-dsDNA antibodies and anti-Sm antibodies; the second correlation coefficient sequence in the correlation array is subjected to quartile calculation to obtain the distribution interval data of anti-Sm antibodies and anti-nRNP antibodies; the kurtosis coefficient and the skewness coefficient of the third correlation coefficient sequence in the correlation array are calculated to obtain the distribution morphology data of anti-dsDNA antibodies and anti-nRNP antibodies; the fluctuation range data, the distribution interval data, and the distribution morphology data are combined to obtain the distribution characteristic data; the correlation array is combined with the distribution characteristic data to obtain the biomarker association feature vector; An evaluation module, used to analyze the dynamic change characteristics of the detection value according to the biomarker-associated feature vector to obtain a biomarker detection evaluation report; An identification module, used to perform feature pattern recognition on abnormal changes in detection values ​​according to the biomarker detection evaluation report to obtain detection data warning information; The analysis module is used to perform dynamic feature analysis on the biomarker detection data based on the detection data warning information to obtain the biomarker detection data analysis results.

Images