A Method for Establishing a Biomarker Association Model for Lupus Erythematosus

Abstract

This invention discloses a method for establishing a biomarker association model for systemic lupus erythematosus (SLE), relating to the field of SLE activity detection technology. The method includes the following steps: acquiring the patient's disease activity score and basic physiological information, and acquiring the patient's first biomarker data and second biomarker data; performing patient labeling and grouping processing to obtain patient labeling group information and patient biomarker data; performing biomarker ranking processing and establishing a biomarker weighting method; constructing a basic activity model based on the patient biomarker data, and constructing an individual activity model for SLE activity detection. This invention addresses the problem that existing technologies for detecting SLE disease activity typically use complex comprehensive disease activity indices, which fail to provide clinicians with a rapid and accurate detection tool to help diagnose and monitor the disease activity of SLE patients.

Description

Technical Field

[0001] This invention relates to the field of SLE activity detection technology, specifically to a method for establishing a biomarker association model for lupus erythematosus. Background Technology

[0002] SLE activity testing technology refers to a technical means of quantitatively assessing the activity level of the disease in patients with systemic lupus erythematosus through a series of laboratory tests and clinical evaluation methods.

[0003] Existing SLE activity assessment technologies typically employ complex comprehensive disease activity indices, such as the SLE Disease Activity Index (SLEDAI) and the Systemic Lupus Erythematosus Activity Measurement (SLAM), to assess SLE activity. This increases the complexity of disease control and poses significant challenges to clinicians in rapid diagnosis and effective management. Alternatively, they may rely solely on the quantitative characteristics of a single biomarker to assess SLE activity. However, SLE is a multi-system autoimmune disease, and focusing on the quantitative characteristics of a single biomarker to determine SLE activity can easily lead to misdiagnosis. There is also the risk of misjudgment, as the detection results of a single biomarker may result in false positives or false negatives. For example, patent application CN112114126A discloses a diagnostic biomarker for systemic lupus erythematosus (SLE) and its application. Although this approach identifies a diagnostic biomarker for SLE, it does not mention the specific detection method for this biomarker. Therefore, existing SLE activity detection technologies typically use complex comprehensive disease activity indices to assess the disease activity of SLE, which cannot provide clinicians with a rapid and accurate detection tool to help them diagnose and monitor the disease activity of SLE patients. Summary of the Invention

[0004] This invention aims to at least partially address one of the technical problems in the prior art. It obtains the patient's disease activity score and basic physiological information, as well as the patient's first and second biomarker data. The patient is then labeled and grouped to obtain patient label grouping information and patient biomarker data. Biomarker ranking is performed, and a biomarker weighting method is established. Based on the patient biomarker data, a basic activity model and an individual activity model are constructed for detecting systemic lupus erythematosus (SLE) activity. This addresses the problem that existing SLE activity detection technologies typically use complex comprehensive disease activity indices for assessment, failing to provide clinicians with a rapid and accurate detection tool to help diagnose and monitor the disease activity of SLE patients.

[0005] To achieve the above objectives, this application provides a method for establishing a biomarker association model for lupus erythematosus, comprising the following steps:

[0006] Obtain the patient's disease activity score and basic physiological information, and obtain the patient's primary biomarker data and secondary biomarker data;

[0007] Patients are labeled and grouped to obtain patient label grouping information and patient biomarker data;

[0008] Based on patient marker grouping information and patient marker data, marker ranking processing is performed, and a marker weighting processing method is established;

[0009] A basic activity model was constructed based on patient biomarker data, and an individual activity model was constructed based on the patient's patient biomarker data. Systemic lupus erythematosus activity was then detected.

[0010] Further, obtaining the patient's disease activity score and basic physiological information, as well as the patient's first biomarker data and second biomarker data, includes the following sub-steps:

[0011] For any patient with systemic lupus erythematosus, denoted as Patient 1, obtain the gender of Patient 1 and record the basic physiological information of Patient 1.

[0012] The systemic lupus erythematosus (SLE) activity score of the first patient was obtained based on the systemic lupus erythematosus disease activity index and recorded as the first patient's disease activity score.

[0013] Furthermore, obtaining the patient's disease activity score and basic physiological information, as well as the patient's primary biomarker data and secondary biomarker data, also includes the following sub-steps:

[0014] The first markers include CD4+ T cells and regulatory T cells; the second markers include CD4+CD39+ T cells and CD39+ regulatory T cells.

[0015] Peripheral blood was collected from the first patient to obtain the percentage and absolute count of CD4+ T cells and the percentage and absolute count of regulatory T cells in the peripheral blood of the first patient, which were recorded as the first biomarker data of the first patient.

[0016] The percentages of CD4+CD39+ T cells and CD39+ regulatory T cells in the peripheral blood of the first patient were obtained and recorded as the second biomarker data of the first patient.

[0017] Repeatedly obtain disease activity scores, basic physiological information, primary biomarker data, and secondary biomarker data from multiple systemic lupus erythematosus patients.

[0018] Further, the patients are labeled and grouped to obtain patient label grouping information and patient biomarker data, including the following sub-steps:

[0019] Based on the gender of the first patient, the first patient is labeled as either a male or female patient;

[0020] The percentage and absolute count of CD4+ T cells, the percentage and absolute count of regulatory T cells, the percentage of CD4+CD39+ T cells, and the percentage of CD39+ regulatory T cells were obtained from all patients with systemic lupus erythematosus (SLE), and marker-based grouping was performed, including:

[0021] For any data of any biomarker, denote it as the first biomarker cell data; arrange the first biomarker cell data of all systemic lupus erythematosus patients in ascending order; denote it as the first biomarker cell sequence; obtain the minimum and maximum values ​​in the first cell sequence, denote them as MX and MD respectively; set the number of sequence groups to k1; calculate the group interval size FD, FD=(MD-MX) / (k1-1).

[0022] Based on the size of the grouping interval FD, the first marker cell sequence is divided into k1 grouping intervals, denoted as [MX-FD / 2, MX+FD / 2), [(MX+FD)-FD / 2, (MX+FD)+FD / 2), [(MX+2*FD)-FD / 2, (MX+2*FD)+FD / 2), ..., [MD-FD / 2, MD+FD / 2]. The k1 grouping intervals are then sequentially denoted as the first grouping interval, the second grouping interval, ..., the k1th grouping interval of the first marker cell data, according to their corresponding order in the first marker cell sequence.

[0023] Furthermore, the process of labeling and grouping patients to obtain patient label grouping information and patient biomarker data includes the following sub-steps:

[0024] For the first patient's first biomarker data and any data of any biomarker from the second biomarker data, label them according to the corresponding grouping interval;

[0025] Repeatedly label the first biomarker data and the second biomarker data of all systemic lupus erythematosus patients. After completion, obtain the first biomarker labeled data and the second biomarker labeled data in sequence, and record them as patient biomarker data.

[0026] For all systemic lupus erythematosus (SLE) patients, those who are of the same sex and whose data for any biomarker fall within the same grouping interval are grouped together and denoted as the first biomarker group, the second biomarker group, ..., the nth biomarker group, and labeled as patient biomarker grouping information.

[0027] Furthermore, the biomarker ranking process based on patient marker grouping information and patient biomarker data, and the establishment of a biomarker weighting method, includes the following sub-steps:

[0028] Set the disease activity score threshold as Q0; any marker group is denoted as any marker group; for any systemic lupus erythematosus (SLE) patient in any marker group, if the corresponding disease activity score is greater than Q0, then the corresponding SLE patient is marked as the active group of any marker group; otherwise, the corresponding SLE patient is marked as the remission group of any marker group. Repeat the marking process for all SLE patients in any marker group.

[0029] Obtain the percentage and absolute count of CD4+ T cells, the percentage and absolute count of regulatory T cells, and the percentage of CD39+ regulatory T cells in all systemic lupus erythematosus patients in the active and remission groups for any marker. Perform ROC curve analysis on any data for any marker and calculate the AUC value. Obtain the AUC values ​​corresponding to the percentage of CD4+ T cells, the absolute count of CD4+ T cells, the percentage of regulatory T cells, the absolute count of regulatory T cells, the percentage of CD4+CD39+ T cells, and the percentage of CD39+ regulatory T cells, and label them as CT1, CT2, TR1, TR2, DD1, and DT1, respectively.

[0030] CT1, CT2, TR1, and TR2 are labeled as the first biomarker relationship parameters, and DD1 and DT1 are labeled as the second biomarker relationship parameters.

[0031] Repeat the process for all marker groups to obtain patient marker grouping information.

[0032] Furthermore, the method for ranking and weighting biomarkers based on patient marker grouping information and patient biomarker data also includes the following sub-steps:

[0033] For the first marker relation parameter of any marker group, calculate MT = CT1 + CT2 + TR1 + TR2; for CT1, CT2, TR1 and TR2, calculate QC1 = CT1 / MT, QC2 = CT2 / MT, QR1 = TR1 / MT and QR2 = TR2 / MT respectively, and denote QC1, QC2, QR1 and QR2 as the marker weights of CT1, CT2, TR1 and TR2 of any marker group respectively;

[0034] For the second marker relation parameter of any marker group, calculate MD=DD1+DT1; for DD1 and DT1, calculate QD1=DD1 / MD and QT1=DT1 / MD respectively; and denote QD1 and QT1 as the marker weights of DD1 and DT1 of any marker group respectively.

[0035] Obtain the marker weights corresponding to all marker groups, and calculate the average values ​​of QC1, QC2, QR1, QR2, QD1 and QT1 respectively, and denot them as PC1, PC2, PR1, PR2, PD1 and PT1 in order.

[0036] Furthermore, a basic activity model is constructed based on patient biomarker data, and an individual activity model is constructed based on the patient's patient biomarker data. The systemic lupus erythematosus activity detection includes the following sub-steps:

[0037] Based on the logistic regression model, a first primitive activity model and a second primitive activity model are constructed. The first primitive activity model is as follows: The second original activity model is as follows: Where g1, g2, g3, g4, h1, and h2 represent, in order, the percentage of CD4+ T cells, the absolute count of CD4+ T cells, the percentage of regulatory T cells, the absolute count of regulatory T cells, the percentage of CD4+CD39+ T cells, and the percentage of CD39+ regulatory T cells; P1 and P2 represent the systemic lupus erythematosus disease activity score; β0, β1, β2, β3, and β4 are the parameters of the first primitive activity model, and α0, α1, and α2 are the parameters of the second primitive activity model;

[0038] The first primitive activity model was trained using the first biomarker data and the corresponding patient's disease activity score, resulting in the first basic activity model. The second primitive activity model was trained using the second biomarker data and the corresponding patient's disease activity score, resulting in the second basic activity model.

[0039] Furthermore, the process of constructing a basic activity model based on patient biomarker data, constructing an individual activity model based on the patient's patient biomarker data, and conducting systemic lupus erythematosus activity detection also includes the following sub-steps:

[0040] For the patient to be tested, acquire the patient's basic physiological information, first biomarker data, and second biomarker data, and acquire the biomarker group corresponding to the patient's basic physiological information, first biomarker data, and second biomarker data, denoted as the first similar group. Acquire the first biomarker data and disease activity score of the patient corresponding to the first similar group, denoted as the first similar data. Acquire the second biomarker data and disease activity score of the patient corresponding to the first similar group, denoted as the second similar data. Acquire QC1, QC2, QR1, QR2, QD1, and QT1 corresponding to the first similar group.

[0041] Furthermore, the process of constructing a basic activity model based on patient biomarker data, constructing an individual activity model based on the patient's patient biomarker data, and conducting systemic lupus erythematosus activity detection also includes the following sub-steps:

[0042] The PC1, PC2, PR1, and PR2 in the first basic activity model are changed in sequence to QC1, QC2, QR1, and QR2 corresponding to the first similar group; and the first basic activity model is trained using the first similar data to obtain the first sexual activity model.

[0043] The PD1 and PT1 in the second basic activity model are changed sequentially to QD1 and QT1 corresponding to the first similar group; and the second basic activity model is trained using the second similar data to obtain the second activity model.

[0044] The first biomarker data and the second biomarker data of the patient to be tested are input into the first sexual activity model and the second sexual activity model in sequence to obtain the first activity A1 and the first activity A2 in sequence; calculate A0=q1*A1+q2*A2 to obtain the systemic lupus erythematosus activity score A0 of the patient to be tested, where q1 and q2 are the set weight coefficients.

[0045] The beneficial effects of this invention are as follows: This invention acquires the patient's disease activity score and basic physiological information, and obtains the patient's first biomarker data and second biomarker data; it performs patient labeling and grouping processing to obtain patient labeling group information and patient biomarker data; based on the patient labeling group information and patient biomarker data, it performs biomarker ranking processing and establishes a biomarker weighting method; it constructs a basic activity model based on the patient biomarker data, and constructs an individual activity model based on the patient's biomarker data, and performs systemic lupus erythematosus (SLE) activity detection; it can provide clinicians with a fast and accurate detection tool to help clinicians diagnose and monitor the disease activity of SLE patients and improve the accuracy of disease activity detection;

[0046] This invention establishes a biomarker weighting method to obtain different weights for different biomarkers and makes the model take into account the importance of the input data. The advantage is that it allows the model to focus more on important biomarker data, thereby more accurately capturing the patterns in the data and improving the accuracy of the model in predicting disease activity. By constructing two activity models with different biomarker data, the advantage is that more appropriate parameter settings can be selected according to the characteristics of the biomarker data they contain. It can more flexibly adapt to the characteristics of different parameter biomarker data combinations, improve the overall performance of the detection model, and thus improve the accuracy of detection. Attached Figure Description

[0047] Figure 1 This is a flowchart of the steps of the method of the present invention;

[0048] Figure 2 This is a flowchart illustrating the first marker cell sequence partitioning process of the present invention;

[0049] Figure 3 This is a schematic diagram of ROC curve analysis according to the present invention;

[0050] Figure 4 This is a scatter plot of the pulse width of the present invention. Detailed Implementation

[0051] 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.

[0052] Example 1, please refer to Figure 1 As shown, in a first aspect, this application provides a method for establishing a biomarker association model for lupus erythematosus, comprising the following steps:

[0053] Step S1 involves obtaining the patient's disease activity score and basic physiological information, as well as the patient's first biomarker data and second biomarker data. Step S1 includes the following sub-steps:

[0054] Step S101: For any patient with systemic lupus erythematosus (SLE), designated as Patient 1, obtain the patient's gender and record it as the patient's basic physiological information; Systemic lupus erythematosus is abbreviated as SLE.

[0055] Step S102: Obtain the systemic lupus erythematosus (SLE) activity score of the first patient based on the SLE disease activity index, and record it as the first patient's disease activity score. The Systemic Lupus Erythematosus Disease Activity Index (SLEDAI) is a commonly used clinical tool for assessing the disease activity of patients with systemic lupus erythematosus. It includes 11 aspects and is scored based on the patient's specific symptoms and examination results, with a total score of 105 points.

[0056] Step S103: The first markers include CD4+ T cells and regulatory T cells; the second markers include CD4+CD39+ T cells and CD39+ regulatory T cells.

[0057] Step S104: Collect peripheral blood from the first patient, obtain the percentage and absolute count of CD4+ T cells and the percentage and absolute count of regulatory T cells in the peripheral blood of the first patient, and record them as the first biomarker data of the first patient; peripheral blood refers to the blood in the circulatory system other than bone marrow, contains a variety of cellular components, has important physiological functions, and has wide applications in clinical diagnosis and other fields.

[0058] Step S105, and obtain the percentage of CD4+CD39+ T cells and the percentage of CD39+ regulatory T cells in the peripheral blood of the first patient, and record them as the second biomarker data of the first patient;

[0059] Step S106: Repeatedly acquire disease activity scores, basic physiological information, first biomarker data, and second biomarker data for multiple systemic lupus erythematosus patients;

[0060] In practical implementation, CD4+ T cells are an important type of immune cell in the human immune system, participating in the pathogenesis of various autoimmune diseases, such as rheumatoid arthritis and systemic lupus erythematosus. Regulatory T cells, or Tregs for short, are a subset of T cells that control the body's autoimmune reactivity. CD39 is an enzyme that plays an important role in the immune system and other physiological processes. As an immunosuppressive molecule, CD39 plays an important role in regulating the immune response in SLE. Incorporating CD39 expression into diagnostic assessment significantly improves the accuracy of determining SLE disease activity. Using CD39+ Treg cells and CD4+CD39+ T cells as biomarkers can provide clinicians with more precise tools to diagnose and monitor disease activity in SLE patients, potentially improving disease management and treatment outcomes.

[0061] Step S2 involves labeling and grouping patients to obtain patient label grouping information and patient biomarker data. Step S2 includes the following sub-steps:

[0062] Step S201: Based on the gender of the first patient, mark the first patient as either a male or female patient. There are significant differences in the clinical manifestations, disease severity, and complications between male and female patients with systemic lupus erythematosus, which are mainly related to gender differences in hormones, genetics, and immune regulation. Therefore, distinguishing between male and female patients facilitates targeted training of the model and improves the accuracy of subsequent activity detection.

[0063] Step S202 involves obtaining the percentage and absolute count of CD4+ T cells, the percentage and absolute count of regulatory T cells, the percentage of CD4+CD39+ T cells, and the percentage of CD39+ regulatory T cells from all systemic lupus erythematosus patients, and performing marker grouping processing. Step S202 includes the following sub-steps:

[0064] For step S2021, please refer to [link / reference]. Figure 2 As shown, any data for any biomarker is denoted as the first biomarker cell data; that is, the percentage and absolute count of CD4+ T cells, the percentage and absolute count of regulatory T cells, the percentage of CD4+CD39+ T cells, and the percentage of CD39+ regulatory T cells. The first biomarker cell data of all systemic lupus erythematosus patients are arranged in ascending order and denoted as the first biomarker cell sequence. The minimum and maximum values ​​in the first cell sequence are obtained and denoted as MX and MD, respectively. The number of sequence groups is set to k1. The size of the grouping interval FD is calculated, FD = (MD-MX) / (k1-1). In this embodiment, k1 = 3, and k1 is generally 2-5.

[0065] Step S2022: Based on the size FD of the grouping interval, divide the first marker cell sequence into k1 grouping intervals, denoted as [MX-FD / 2, MX+FD / 2), [(MX+FD)-FD / 2, (MX+FD)+FD / 2), [(MX+2*FD)-FD / 2, (MX+2*FD)+FD / 2), ..., [MD-FD / 2, MD+FD / 2]. And according to the order of the k1 grouping intervals in the first marker cell sequence, denoted as the first grouping interval, the second grouping interval, ..., the k1th grouping interval of the first marker cell data.

[0066] Step S2023: For the first biomarker data and any data of any biomarker in the second biomarker data of the first patient, label them according to the corresponding grouping interval;

[0067] Step S2024: Repeat the labeling of the first biomarker data and the second biomarker data for all patients with systemic lupus erythematosus. After completion, obtain the first biomarker labeling data and the second biomarker labeling data in sequence, and record them as patient biomarker data.

[0068] Step S203: For all systemic lupus erythematosus (SLE) patients, those who are of the same sex and whose data for any biomarker fall within the same grouping interval are divided into a group, which is denoted as the first biomarker group, the second biomarker group, ... the nth biomarker group, and marked as patient biomarker group information;

[0069] In practice, k1 should not be too large. If it is too large, there will be fewer patients in the same group, that is, less data in the same group, which will result in insufficient data for subsequent model training. Theoretically, the number of groups is 2*k1^6. However, due to the certain variation relationship between the six markers of CD4+ T cell percentage and absolute count, regulatory T cell percentage and absolute count, CD4+CD39+ T cell percentage and CD39+ regulatory T cell percentage, the actual number of groups will be much less than 2*k1^6, thus ensuring the amount of data in the same group.

[0070] Step S3 involves sorting the biomarkers based on patient marker grouping information and patient biomarker data, and establishing a biomarker weighting method. Step S3 includes the following sub-steps:

[0071] Step S301: Set the disease activity score threshold to Q0; any marker group is denoted as any marker group; for any systemic lupus erythematosus (SLE) patient within any marker group, if the corresponding disease activity score is greater than Q0, the corresponding SLE patient is marked as the active group of any marker group; otherwise, the corresponding SLE patient is marked as the remission group of any marker group. This marking process is repeated for all SLE patients within any marker group. In this embodiment, Q0=9, that is, patients with a disease activity score greater than 9 are marked as the active group, and patients with a disease activity score of 9 or less are marked as the remission group.

[0072] For step S302, please refer to... Figure 3 As shown, Figure 3 In the graph, A, B, and C are, in order, the percentage graph of CD39+ regulatory T cells, the percentage graph of regulatory T cells, and the ROC curve analysis graph of the absolute count of regulatory T cells, respectively. D is a combined graph of the curves of A, B, and C. Figure 3E, F, and G in the figure represent, in order, the percentage of CD4+CD39+ T cells, the percentage of CD4+ T cells, and the ROC curve analysis of the absolute count of CD4+ T cells; H is the combined curve of E, F, and G; the percentage and absolute count of CD4+ T cells, the percentage and absolute count of regulatory T cells, and the percentage of CD39+ regulatory T cells are obtained for all systemic lupus erythematosus patients in the active and remission groups for any marker; ROC curve analysis is performed on any data for any marker, and the AUC value is calculated to obtain the AUC values ​​corresponding to the percentage of CD4+ T cells, the absolute count of CD4+ T cells, the percentage of regulatory T cells, the absolute count of regulatory T cells, the percentage of CD4+CD39+ T cells, and the percentage of CD39+ regulatory T cells, respectively, and are denoted as CT1, CT2, TR1, TR2, DD1, and DT1 in sequence;

[0073] ROC curve analysis is a statistical method used to evaluate the performance of binary classification models. The ROC curve plots the false positive rate on the x-axis and the true positive rate on the y-axis. In binary classification problems, the model makes a prediction for each sample, giving a probability value for predicting it as positive. By setting different probability thresholds, samples can be divided into positive and negative classes. The AUC value, which is the area under the ROC curve, is an important indicator of ROC curve performance. The AUC value ranges from 0.5 to 1.

[0074] Step S303: Mark CT1, CT2, TR1 and TR2 as the first marker relationship parameters, and mark DD1 and DT1 as the second marker relationship parameters;

[0075] Step S304: Repeat the processing of all marker groups to obtain patient marker grouping information;

[0076] Step S305: For the first marker relationship parameter of any marker group, calculate MT = CT1 + CT2 + TR1 + TR2; for CT1, CT2, TR1, and TR2, calculate QC1 = CT1 / MT, QC2 = CT2 / MT, QR1 = TR1 / MT, and QR2 = TR2 / MT respectively, and denote QC1, QC2, QR1, and QR2 as the marker weights of CT1, CT2, TR1, and TR2 of any marker group; for example, if CT1 = 0.56, CT2 = 0.53, TR1 = 0.64, and TR2 = 0.52, then MT = 2.25, QC1 = 0.56 / 2.25 = 0.25, QC2 = 0.53 / 2.25 = 0.24, QR1 = 0.64 / 2.25 = 0.28, and QR2 = 0.52 / 2.25 = 0.23;

[0077] Step S306: For the second marker relationship parameter of any marker group, calculate MD = DD1 + DT1; for DD1 and DT1, calculate QD1 = DD1 / MD and QT1 = DT1 / MD respectively; record QD1 and QT1 as the marker weights of DD1 and DT1 of any marker group; for example, if DD1 = 0.71 and DT1 = 0.75, then MD = 1.46, QC1 = 0.71 / 1.46 = 0.49, and QC2 = 0.75 / 1.46 = 0.51.

[0078] Step S307: Obtain the marker weights corresponding to all marker groups, and calculate the average values ​​of QC1, QC2, QR1, QR2, QD1 and QT1 respectively, and record them as PC1, PC2, PR1, PR2, PD1 and PT1 in order.

[0079] In practice, the AUC value provides a quantitative indicator that accurately reflects the ability of each biomarker to distinguish different patient groups. The AUC value ranges from 0.5 to 1. The closer the value is to 1, the stronger the predictive ability of the biomarker for grouping, and the higher its importance, i.e., the greater its weight. The closer the value is to 0.5, the weaker the predictive ability of the parameter, and the relatively lower its importance, i.e., the lower its weight. Through this quantitative method, the differences in importance between different biomarker data can be clearly compared.

[0080] Step S4 involves constructing a basic activity model based on patient biomarker data, constructing an individualized activity model based on the patient's patient biomarker data, and performing systemic lupus erythematosus activity detection. Step S4 includes the following sub-steps:

[0081] Step S401: Construct a first primitive activity model and a second primitive activity model based on the logistic regression model. The first primitive activity model is as follows: The second original activity model is as follows: Where g1, g2, g3, g4, h1, and h2 represent, in order, the percentage of CD4+ T cells, the absolute count of CD4+ T cells, the percentage of regulatory T cells, the absolute count of regulatory T cells, the percentage of CD4+CD39+ T cells, and the percentage of CD39+ regulatory T cells; P1 and P2 represent the systemic lupus erythematosus disease activity score; β0, β1, β2, β3, and β4 are the parameters of the first original activity model, and α0, α1, and α2 are the parameters of the second original activity model; before model training, β0, β1, β2, β3, β4, α0, α1, and α2 need to be initialized. Generally, β0, β1, β2, β3, β4, α0, α1, and α2 are initialized to small random values, for example, 0.004;

[0082] Step S402: The first basic activity model is trained using the first biomarker data and the corresponding patient's disease activity score, resulting in the first original activity model. Then, the second basic activity model is trained using the second biomarker data and the corresponding patient's disease activity score, resulting in the second original activity model. This training uses all the corresponding biomarker data, allowing the model to access as many samples and features as possible, thereby learning the widely existing patterns and rules in the data. This enables the model to possess more comprehensive knowledge, a more general understanding and adaptability to various situations, and avoids the model focusing only on local data features and making biased judgments.

[0083] Step S403: For the patient to be tested, acquire the patient's basic physiological information, first biomarker data, and second biomarker data, and acquire the biomarker group corresponding to the patient's basic physiological information, first biomarker data, and second biomarker data, denoted as the first similar group; acquire the first biomarker data and disease activity score of the patient corresponding to the first similar group, denoted as the first similar data; acquire the second biomarker data and disease activity score of the patient corresponding to the first similar group, denoted as the second similar data; and acquire QC1, QC2, QR1, QR2, QD1, and QT1 corresponding to the first similar group; for example, if the patient to be tested is male, and his / her gender plus the first biomarker data and the second biomarker data are located in the second biomarker group, then the second biomarker group is denoted as the first similar group;

[0084] Step S404: PC1, PC2, PR1 and PR2 in the first basic activity model are changed in sequence to QC1, QC2, QR1 and QR2 corresponding to the first similar group; and the first basic activity model is trained using the first similar data to obtain the first sexual activity model.

[0085] Step S405: Change PD1 and PT1 in the second basic activity model to QD1 and QT1 corresponding to the first similar group in sequence; and use the second similar data to train the second basic activity model to obtain the second activity model.

[0086] This training utilizes data from patients similar to the test patient, allowing the model to be further optimized for the characteristics of a specific patient group. Because similar patients may have similar disease characteristics, physiological conditions, or other related factors, by focusing on this part of the data, the model can learn more deeply about specific patterns and characteristics related to this patient group, thereby improving the accuracy and targeting of predictions for this group of patients.

[0087] Step S406: Input the first biomarker data and the second biomarker data of the patient to be tested into the first sexual activity model and the second sexual activity model in sequence to obtain the first activity A1 and the first activity A2 in sequence; calculate A0=q1*A1+q2*A2 to obtain the systemic lupus erythematosus activity score A0 of the patient to be tested, where q1 and q2 are set weight coefficients; in this embodiment, q1 is 0.4 and q2 is 0.6, which can be set according to the actual application scenario;

[0088] In practical implementation, different weights are assigned to different biomarkers, and the model takes into account the importance of the input data. Weights can be allocated based on the actual contribution of each biomarker to disease activity. For example, if the percentage of CD39+ regulatory T cells has a significant impact on activity in practice, giving it a higher weight allows the model to focus more on this parameter, thereby more accurately capturing patterns in the data and improving the model's accuracy in predicting disease activity. It also avoids the model from over-relying on some less important data or being affected by noisy data. When the weight of a certain data is set correctly, even if there are some small fluctuations in the data of that parameter, it will not have a significant impact on the model's output, thus making the model more stable and reducing the uncertainty of the model's output.

[0089] Example 2, Second Aspect, Please refer to Figure 4 As shown, Figure 4 A schematic diagram of an electronic device is provided, which may include a processor, a communication interface, a memory, and a communication bus. The processor, communication interface, and memory communicate with each other via the communication bus. The memory stores computer-readable instructions, and the processor can call these instructions. When the processor executes a computer-readable instruction, it performs steps as described in a method for establishing a biomarker association model for lupus erythematosus, to achieve the following functions: obtaining the patient's disease activity score and basic physiological information, and obtaining the patient's first biomarker data and second biomarker data; performing patient labeling and grouping processing to obtain patient labeling grouping information and patient biomarker data; performing biomarker ranking processing based on the patient labeling grouping information and patient biomarker data, and establishing a biomarker weighting method; constructing a basic activity model based on the patient biomarker data, and constructing an individual activity model based on the patient's patient biomarker data, and performing systemic lupus erythematosus activity detection.

[0090] Furthermore, when the logical instructions in the aforementioned memory can be implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0091] In a third aspect, this application also provides a computer-readable storage medium storing a computer program. When executed by a processor, the computer program performs the steps described above in the method for establishing a biomarker association model for lupus erythematosus, to achieve the following functions: acquiring the patient's disease activity score and basic physiological information, and acquiring the patient's first biomarker data and second biomarker data; performing patient labeling and grouping processing on the patient to obtain patient labeling grouping information and patient biomarker data; performing biomarker ranking processing based on the patient labeling grouping information and patient biomarker data, and establishing a biomarker weighting processing method; constructing a basic activity model based on the patient biomarker data, and constructing an individual activity model based on the patient's patient biomarker data, and performing systemic lupus erythematosus activity detection.

[0092] Based on the above description of the embodiments, the embodiments of the present invention can be provided as methods, systems, or computer program products. Based on this understanding, the above technical solutions, in essence or in terms of their contribution to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or certain parts of the embodiments.

[0093] In the embodiments provided in this application, it should be understood that the disclosed system or method can be implemented in other ways. The embodiments described above are merely illustrative. For example, the division of modules or units is only a logical functional division, and there may be other division methods in actual implementation. Furthermore, multiple modules or units may be combined or integrated into another system, or some features may be ignored or not executed. Additionally, the coupling or direct coupling or communication connection shown or discussed may be through some communication interfaces. The indirect coupling or communication connection between systems, modules, and units may be electrical, mechanical, or other forms.

[0094] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A method for establishing a biomarker association model for lupus erythematosus, characterized in that, The steps include: The study obtained the patient's disease activity score and basic physiological information, as well as the patient's primary and secondary biomarker data. The basic physiological information included the patient's gender. The primary biomarker data included the percentage and absolute count of CD4+ T cells and the percentage and absolute count of regulatory T cells in the patient's peripheral blood. The secondary biomarker data included the percentage of CD4+CD39+ T cells and the percentage of CD39+ regulatory T cells in the patient's peripheral blood. Patients were then labeled and grouped to obtain patient labeling information and patient biomarker data. Based on patient marker grouping information and patient marker data, marker ranking processing is performed, and a marker weighting processing method is established; A basic activity model was constructed based on patient biomarker data, and an individual activity model was constructed based on the patient's patient biomarker data. Systemic lupus erythematosus activity was then detected. The process of labeling and grouping patients to obtain patient label grouping information and patient biomarker data includes the following sub-steps: For any patient with systemic lupus erythematosus, we will refer to them as Patient 1. Based on the gender of Patient 1, we will label Patient 1 as either male or female. The percentage and absolute count of CD4+ T cells, the percentage and absolute count of regulatory T cells, the percentage of CD4+CD39+ T cells, and the percentage of CD39+ regulatory T cells were obtained for all systemic lupus erythematosus (SLE) patients. Patient labeling and grouping were then performed, including: For any data of any biomarker, denote it as the first biomarker cell data; arrange the first biomarker cell data of all systemic lupus erythematosus patients in ascending order, and denote it as the first biomarker cell sequence; obtain the minimum and maximum values ​​in the first biomarker cell sequence, and denote them as MX and MD respectively; set the number of sequence groups to k1; calculate the group interval size FD, FD=(MD-MX) / (k1-1). Based on the size of the grouping interval FD, the first marker cell sequence is divided into k1 grouping intervals, denoted as [MX-FD / 2, MX+FD / 2], [(MX+FD)-FD / 2, (MX+FD)+FD / 2], [(MX+2*FD)-FD / 2, (MX+2*FD)+FD / 2], ..., [MD-FD / 2, MD+FD / 2]. These k1 grouping intervals are then sequentially denoted as the first grouping interval, the second grouping interval, ..., the k1th grouping interval of the first marker cell data, according to their corresponding order in the first marker cell sequence. For the first patient's first biomarker data and any data of any biomarker from the second biomarker data, label them according to the corresponding grouping interval; Repeatedly label the first biomarker data and the second biomarker data of all systemic lupus erythematosus patients. After completion, obtain the first biomarker labeled data and the second biomarker labeled data in sequence, and record them as patient biomarker data. For all systemic lupus erythematosus (SLE) patients, those who are of the same sex and whose data for any biomarker fall in the same grouping interval are grouped together and denoted as the first biomarker group, the second biomarker group, ..., the nth biomarker group, and labeled as patient biomarker group information; The method for ranking and weighting biomarkers based on patient marker grouping information and patient biomarker data includes the following sub-steps: Set the disease activity score threshold as Q0; any marker group is denoted as any marker group; for any systemic lupus erythematosus (SLE) patient in any marker group, if the corresponding disease activity score is greater than Q0, then the corresponding SLE patient is marked as the active group of any marker group; otherwise, the corresponding SLE patient is marked as the remission group of any marker group. Repeat the marking process for all SLE patients in any marker group. Obtain the percentage and absolute count of CD4+ T cells, the percentage and absolute count of regulatory T cells, the percentage of CD4+CD39+ T cells, and the percentage of CD39+ regulatory T cells in all systemic lupus erythematosus patients in the active and remission groups for any biomarker. Perform ROC curve analysis on any data for any biomarker and calculate the AUC value. Obtain the AUC values ​​corresponding to the percentage of CD4+ T cells, the absolute count of CD4+ T cells, the percentage of regulatory T cells, the absolute count of regulatory T cells, the percentage of CD4+CD39+ T cells, and the percentage of CD39+ regulatory T cells, and label them as CT1, CT2, TR1, TR2, DD1, and DT1, respectively. CT1, CT2, TR1, and TR2 are labeled as the first biomarker relationship parameters, and DD1 and DT1 are labeled as the second biomarker relationship parameters. Repeat the process for all marker groups to obtain patient marker grouping information. For the first marker relation parameter of any marker group, calculate MT = CT1 + CT2 + TR1 + TR2; for CT1, CT2, TR1 and TR2, calculate QC1 = CT1 / MT, QC2 = CT2 / MT, QR1 = TR1 / MT and QR2 = TR2 / MT respectively, and denote QC1, QC2, QR1 and QR2 as the marker weights of CT1, CT2, TR1 and TR2 of any marker group respectively; For the second marker relation parameter of any marker group, calculate MD=DD1+DT1; for DD1 and DT1, calculate QD1=DD1 / MD and QT1=DT1 / MD respectively; and denote QD1 and QT1 as the marker weights of DD1 and DT1 of any marker group respectively. Obtain the marker weights corresponding to all marker groups, and calculate the average values ​​of QC1, QC2, QR1, QR2, QD1 and QT1 respectively, and denot them as PC1, PC2, PR1, PR2, PD1 and PT1 in order; The process of constructing a basic activity model based on patient biomarker data, constructing an individual activity model based on patient biomarker data, and performing systemic lupus erythematosus activity detection includes the following sub-steps: Based on the logistic regression model, a first primitive activity model and a second primitive activity model are constructed. The first primitive activity model is as follows: The second original activity model is as follows: Where g1, g2, g3, g4, h1, and h2 represent, in order, the percentage of CD4+ T cells, the absolute count of CD4+ T cells, the percentage of regulatory T cells, the absolute count of regulatory T cells, the percentage of CD4+CD39+ T cells, and the percentage of CD39+ regulatory T cells; P1 and P2 represent the systemic lupus erythematosus disease activity score; β0, β1, β2, β3, and β4 are the parameters of the first primitive activity model, and α0, α1, and α2 are the parameters of the second primitive activity model; The first primitive activity model was trained using the first biomarker data and the corresponding patient's disease activity score to obtain the first basic activity model. The second primitive activity model was trained using the second biomarker data and the corresponding patient's disease activity score to obtain the second basic activity model. For the patient to be tested, acquire the patient's basic physiological information, first biomarker data, and second biomarker data, and acquire the biomarker group corresponding to the patient's basic physiological information, first biomarker data, and second biomarker data, denoted as the first similar group. Acquire the first biomarker data and disease activity score of the patient corresponding to the first similar group, denoted as the first similar data; acquire the second biomarker data and disease activity score of the patient corresponding to the first similar group, denoted as the second similar data; and acquire QC1, QC2, QR1, QR2, QD1, and QT1 corresponding to the first similar group. The PC1, PC2, PR1, and PR2 in the first basic activity model are changed in sequence to QC1, QC2, QR1, and QR2 corresponding to the first similar group; and the modified first basic activity model is trained using the first similar data to obtain the first sexual activity model. The PD1 and PT1 in the second basic activity model are changed sequentially to QD1 and QT1 corresponding to the first similar group; and the second similar data is used to train the modified second basic activity model to obtain the second activity model. The first biomarker data and the second biomarker data of the patient to be tested are input into the first sexual activity model and the second sexual activity model in sequence, respectively, to obtain the first activity A1 and the second activity A2 in sequence; calculate A0=q1*A1+q2*A2 to obtain the systemic lupus erythematosus disease activity score A0 of the patient to be tested, where q1 and q2 are the set weight coefficients.

2. The method for establishing a biomarker association model for lupus erythematosus according to claim 1, characterized in that, Obtaining the patient's disease activity score and basic physiological information, as well as the patient's primary biomarker data and secondary biomarker data, includes the following sub-steps: The systemic lupus erythematosus (SLE) activity score of the first patient was obtained based on the systemic lupus erythematosus disease activity index and recorded as the first patient's disease activity score.

Images