Combinations of protein markers and clinical indicators, products and uses thereof

Abstract

The application discloses a combination of protein markers and clinical indexes related to leukoaraiosis, products and applications thereof. The protein marker combination comprises P05198, P00966, Q8IZP0 and P48061. The clinical indexes comprise MMSE, CTTtrail1 and SDMT. The application also provides applications, kits and computer program products of the combination. The combination can be used for quickly, conveniently and economically identifying leukoaraiosis. The computer program product has good model performance and shows relatively accurate prediction ability, and has clinical use and promotion value.

Description

Technical Field

[0001] This invention belongs to the technical field of combining biomedicine and machine learning, specifically involving the combination of protein biomarkers and clinical indicators related to brain white matter lesions, products and their applications. Background Technology

[0002] White matter lesions (WML) are a type of brain pathology closely related to vascular risk factors and aging, characterized by demyelination, axonal degeneration, and small vessel disease in the white matter region. WML is a significant risk factor for neurodegenerative diseases such as vascular cognitive impairment, stroke, Alzheimer's disease, and Parkinson's disease. WML is primarily detected using T2-weighted imaging or fluid attenuated inversion recovery (FLAIR) sequences of the head, presenting as high-signal punctate, patchy, or confluent lesions. The Fazekas score is commonly used to grade the extent of damage in the detection of white matter lesions. WML is common in middle-aged and elderly individuals; over 87% of people aged 60–70 years have WML, and the detection rate is as high as 95%–100% in people aged 80–90 years. Numerous studies have shown that vascular risk factors such as hypertension, diabetes, hyperlipidemia, coronary heart disease, and smoking can promote the occurrence of WML and accelerate its progression. However, the severity of white matter damage varies considerably among individuals who have been exposed to these vascular risk factors for a long period of time, suggesting that individual susceptibility to vascular risk factors may be influenced by other biological factors.

[0003] Although white matter damage is common in middle-aged and elderly individuals with long-term exposure to vascular risk factors, the degree of damage varies significantly among individuals. This heterogeneity may be related to multiple factors, including genetic factors, lifestyle, and environmental exposure. Therefore, relying solely on traditional clinical indicators is insufficient to accurately assess and predict the extent of white matter damage. Currently, the assessment of white matter damage primarily relies on head magnetic resonance imaging (MRI), with the Fazekas score being a commonly used semi-quantitative assessment method. The Fazekas score classifies white matter damage into mild (score < 2) and severe (score ≥ 2).

[0004] However, the high cost, long scanning time, and limited equipment accessibility of MRI examinations restrict the feasibility of large-scale screening and early identification. Therefore, a more convenient and economical biomarker is needed to assist in the assessment of white matter damage. On the other hand, although traditional hematological indicators (such as blood lipids, blood glucose, and inflammatory markers) are related to white matter lesions (WML), the predictive power of a single indicator is limited and cannot accurately reflect the pathological changes of WML in an individual. Therefore, exploring new biomarkers to establish a simple, economical, and scalable blood testing model is of great clinical significance for the early screening and disease management of WML. Summary of the Invention

[0005] To address the aforementioned technical issues, this application provides a combination of protein biomarkers and clinical indicators related to white matter lesions, products, and their applications.

[0006] The technical solution provided in this application is as follows:

[0007] In a first aspect, this application provides a combination of protein biomarkers, including: P05198, P00966, Q8IZP0 and P48061.

[0008] Secondly, this application provides the application of a reagent for detecting a combination of protein biomarkers in the preparation of products for diagnosing white matter lesions, wherein the combination of protein biomarkers includes P05198, P00966, Q8IZP0 and P48061.

[0009] Thirdly, this application provides a combination of protein biomarkers and clinical indicators, wherein the protein biomarkers include: P05198, P00966, Q8IZP0 and P48061; and the clinical indicators include MMSE, CTTtrail1 and SDMT.

[0010] Fourthly, this application provides the application of a reagent for detecting a combination of protein biomarkers and clinical indicators in the preparation of products for diagnosing leukoencephalopathy, wherein the combination of protein biomarkers includes P05198, P00966, Q8IZP0, and P48061; and the clinical indicators include MMSE, CTTtrail1, and SDMT.

[0011] Fifthly, this application provides a kit comprising detection reagents for detecting a combination of the protein biomarkers and clinical indicators described in the third aspect.

[0012] Sixthly, this application provides the use of the kit described in the fifth aspect in the preparation of products for detecting leukoencephalopathy, and the use of the detection reagent in the kit described in the fifth aspect in the preparation of kits for diagnosing leukoencephalopathy.

[0013] In a seventh aspect, this application provides a program product related to leukoencephalopathy, the computer program product being used to diagnose the risk of a subject having leukoencephalopathy, including the following steps:

[0014] The expression levels of each protein biomarker in the plasma of the subject were obtained; the protein biomarkers included P05198, P00966, Q8IZP0, and P48061.

[0015] Substitute the expression levels of each protein biomarker and the values ​​of clinical indicators into the binary logistic regression equation to calculate the logarithm y of the advantage of the subject.

[0016] Calculate the probability P that the subject is a healthy person based on y, P = exp(y) / {1 + exp(y)}, where exp(y) represents the natural exponential function;

[0017] Based on the comparison between probability P and reference value, the test subject is diagnosed or predicted to have leukoencephalopathy or to be at risk of having leukoencephalopathy.

[0018] In one possible implementation, the formula for the binary logistic regression equation is:

[0019]

[0020] Where A is the intercept term, B1 to B4 are the regression coefficients of the independent variables, and x1, x2, x3, and x4 are the expression levels of proteins P05198, P00966, Q8IZP0, and P48061, respectively.

[0021] Furthermore, the method for obtaining the binary logistic regression equation includes:

[0022] To obtain the expression levels of protein markers in healthy individuals and patients with white matter lesions;

[0023] Using the expression level of protein markers as input and whether or not the disease is present as output, a binary logistic regression equation is obtained through training.

[0024] Eighthly, this application provides a program product related to leukoencephalopathy, the computer program product being used to diagnose the risk of a subject having leukoencephalopathy, including the following steps:

[0025] The expression levels of each protein biomarker and the values ​​of each clinical indicator in the plasma of the test subjects were obtained; the protein biomarkers included P05198, P00966, Q8IZP0 and P48061; the clinical indicators included MMSE, CTTtrail1 and SDMT.

[0026] Substitute the expression levels of each protein biomarker and the values ​​of clinical indicators into the binary logistic regression equation to calculate the logarithm y of the advantage of the subject.

[0027] Calculate the probability P that the subject is a healthy person based on y, P = exp(y) / {1 + exp(y)}, where exp(y) represents the natural exponential function;

[0028] Based on the comparison between probability P and reference value, the test subject is diagnosed or predicted to have leukoencephalopathy or to be at risk of having leukoencephalopathy.

[0029] In one possible implementation, the formula for the binary logistic regression equation is:

[0030]

[0031] Where A is the intercept term, B1 to B7 are the regression coefficients of the independent variables; x1, x2, x3, and x4 are the expression levels of proteins P05198, P00966, Q8IZP0, and P48061, respectively; and x5, x6, and x7 are the expression levels of clinical indicators SDMT, MMSE, and CTTtrail1, respectively.

[0032] Furthermore, the method for obtaining the binary logistic regression equation includes:

[0033] To obtain the expression levels of protein markers and the values ​​of clinical indicators in healthy individuals and patients with white matter lesions;

[0034] Using the expression levels of protein biomarkers and the values ​​of clinical indicators as inputs, and whether or not a disease is present as output, a binary logistic regression equation is obtained through training.

[0035] Ninthly, this application provides a method for constructing a combined model of protein biomarkers and clinical indicators, including:

[0036] To obtain clinical indicators and plasma protein expression data of patients with leukoencephalopathy;

[0037] By comparing protein expression levels with those of healthy individuals, the data of the first differentially expressed proteins were selected.

[0038] The importance scores and rankings of differentially expressed proteins in distinguishing different groups of samples were calculated using the XGBoost machine learning model and the LASSO regression model, and the data of the second differentially expressed protein were screened out.

[0039] The second differentially expressed protein data and clinical indicator information were randomly combined as input, and whether or not the disease was the output. Multiple combined models were constructed using logistic regression.

[0040] The optimal combination model is selected based on the AUC value.

[0041] The beneficial effects of this application are as follows:

[0042] (1) This application provides a combination, product and application of protein biomarkers and clinical indicators for assessing the degree of white matter lesions that do not rely entirely on clinical scales and neuroimaging, which can be used to quickly, conveniently and economically identify white matter lesions.

[0043] (2) This application also provides a computer program product that can construct a predictive model using a combination of protein biomarkers and clinical indicators to effectively assess the degree of white matter lesions. The predictive model has good performance, showing relatively accurate predictive ability, and has clinical use and promotion value. Attached Figure Description

[0044] Figure 1 This is a statistical count of the number of proteins quantified in Example 1;

[0045] Figure 2 For screening of differentially expressed proteins in Example 1;

[0046] Figure 3 These are the prediction model parameters for the protein biomarker combinations in Example 1;

[0047] Figure 4 These are the predictive model parameters for the combination of protein biomarkers and clinical indicators in Example 1;

[0048] Figure 5 Nodal plot for protein biomarker combinatorial models;

[0049] Figure 6 ROC analysis plot for predictive models of protein biomarker combinations;

[0050] Figure 7 Confusion matrix diagram of a prediction model for a combination of protein biomarkers;

[0051] Figure 8 Box plot of scores for a predictive model of protein biomarker combinations;

[0052] Figure 9 Box plot of scores for a predictive model combining protein biomarkers and clinical indicators;

[0053] Figure 10 ROC analysis plot for a predictive model combining protein biomarkers and clinical indicators;

[0054] Figure 11 Confusion matrix analysis of a predictive model combining protein biomarkers and clinical indicators. Detailed Implementation

[0055] The content of this application will be further described below with reference to specific embodiments, but the content of this application is not limited thereto.

[0056] Currently, the high cost, long scanning time, and limited equipment accessibility of MRI examinations restrict the feasibility of large-scale screening and early identification.

[0057] In view of this, this application provides a combination of protein biomarkers and clinical indicators related to white matter lesions, products and their applications.

[0058] For ease of understanding, the relevant technical terms appearing in the embodiments are explained uniformly.

[0059] Protein biomarkers:

[0060] P05198: Gene name EIF2S1, protein name Eukaryotic translation initiation factor2 subunit 1;

[0061] P00966: Gene name ASS1, protein name Argininosuccinate synthase;

[0062] Q8IZP0: Gene name ABI1, protein name Abl interactor 1;

[0063] P48061: Gene name CXCL12, protein name Stromal cell-derived factor 1.

[0064] Clinical indicators:

[0065] CI: Cognitive Impairment

[0066] SDMT: Symbolic Number Conversion Test

[0067] MMSE: Mini-Mental State Examination

[0068] HLP: Hyperlipoproteinemia

[0069] education_year: Years of education

[0070] MoCA: Montreal Cognitive Assessment Scale

[0071] CTT_Trail1: Color Trail Test, Number Color Connection Test 1

[0072] smoke_status: Smoking status

[0073] CTT_Trail2: Color Trail Test, Number Color Connection Test 2

[0074] sex: gender

[0075] CDT3: Clock Drawing Test (3 points)

[0076] TMT_B: Wiring Test B

[0077] TMT_A: Wiring Test A

[0078] BNT_15: Boston Naming Test (15 items)

[0079] Coronary artery disease (CHD)

[0080] AF: Atrial fibrillation

[0081] VHD: Valvular Heart Disease

[0082] age: age

[0083] HAMA: Hamilton Anxiety Rating Scale

[0084] HAMD: Hamilton Depression Rating Scale

[0085] diabetes

[0086] drink_status: Drinking status

[0087] BMI: Body Mass Index, calculated as weight divided by the square of height.

[0088] MI: Myocardial infarction

[0089] hypertension: high blood pressure

[0090] TIA: Transient ischemic attack

[0091] PVD: Peripheral Vascular Disease

[0092] HCY: Hyperhomocysteinemia

[0093] Example 1: Screening of biomarkers for brain white matter lesions and construction of a diagnostic model

[0094] 1. Sample Information

[0095] The samples used in this embodiment were obtained from 181 patients at Zhongnan Hospital of Wuhan University. Patient information is shown in Table 1.

[0096] Table 1 Patient Clinical Information Form

[0097]

[0098] Note: Group0: Healthy control group; Group1: Leukoencephalopathy group; sex: 0 represents male, 1 represents female; smoke_status: 0 represents currently non-smoker, 1 represents smoker; drink_status: 0 represents currently non-drinker, 1 represents drinker; hypertension, diabetes, HLP, MI, CHD, AF, VHD, PVD, TIA, CI, HCY: 0 represents not having the disease, 1 represents having the disease; CDT3: 0, 1, 2, 3 represent scores respectively; Categorical variable information, such as in the sex (%) column, the value represents the number of patients in each category, such as 70 means there are 70 males among all patients; Continuous variable information, such as in the age (median [IQR]) column, the value represents the median, such as 65.00 means the median age of all patients is 65; Other numerical data are the same.

[0099] 2. Sample pretreatment

[0100] Take 20 μL of plasma sample, dilute with loading buffer (10 mM Tris-Cl, 1 mM EDTA, 150 mM KCl, 0.05% CHAPS), mix thoroughly with 1 mg of superparamagnetic iron oxide nanobead suspension, and incubate at 37 °C for 1 hour. Wash the nanobeads twice with loading buffer, and then wash them once with loading buffer (10 mM Tris-Cl, 1 mM EDTA, 150 mM KCl) without CHAPS [(3-[(3-cholamidopropyl)dimethylamino]-1-propanesulfonic acid inner salt)]. Adsorb the nanobeads onto a magnetic rack, discard the supernatant, and obtain the protein-enriched nanobeads. Add lysis buffer (1% SDC / 100 mM Tris-HCl, pH=8.5 / 10 mM TCEP / 40 mM CAA) to the sample, and incubate at 60 °C for 30 min to perform a reductive alkylation reaction. Add an equal volume of ddH2O to dilute the SDC (sodium deoxycholate) concentration to below 0.5%, add 1 μg of trypsin, and incubate overnight at 37 °C with shaking for enzyme digestion. On the second day, add TFA to terminate the enzyme digestion, collect the supernatant for desalting on an SDB-RPS desalting column, vacuum dry, and store at -20 °C for later use.

[0101] 3. LC-MS / MS detection and analysis

[0102] Mass spectrometry analysis of the samples was performed using an UltiMate 3000 RSLC nano-liquid chromatography (Thermo) tandem timsTOFPro mass spectrometer (Bruker). Peptide samples were injected via autosampler and bound to a C18 trap column (75 µm × 2 cm, 3 µm particle size, 100 Å pore size, Thermo), followed by separation in an analytical column (75 µm × 15 cm, 1.7 µm particle size, 100 Å pore size, IonOpticks). Analytical gradients were established using mobile phase A (0.1% formicacid) and mobile phase B (0.1% formic acid in ACN). Mass spectrometry data acquisition was performed in diaPASEF mode. The capillary voltage was set to 1500 V. The scan range for MS1 and MS2 spectra was set to 100–1700 m / z. The ion mobility range was set to 0.6–1.6 Vs / cm. 2 Accumulation time and ramp time were set to 50 ms. Based on the mass-to-charge ratio-ion mobility distribution, the diaPASEF acquisition window was set using timsControl software. The collision energy was set to 1 / K0 = 1.6 Vs / cm² based on the ion mobility. 2 The voltage decreased linearly from 59 eV to 20 eV (1 / K0 = 0.6 Vs / cm²), thus obtaining the raw DIA data file for the mass spectrometer.

[0103] 4. Data Preprocessing

[0104] Raw DIA data files were analyzed using DIA-NN software (v 1.8). The database used for searching was the Human proteome reference database in Uniprot (February 9, 2022, containing 20375 protein sequences). A spectral library was predicted using a deep learning algorithm in DIA-NN. The predicted spectral library and the spectral library obtained through the MBR function were used to extract from the raw DIA data, and protein expression quantification values ​​were calculated based on mass spectrometry detection signals. The final results were filtered for precursor ions and protein levels using a 1% FDR. The filtered proteome quantification information was used for subsequent analysis. Figure 1 As shown, 957-2912 different proteins were quantified in each sample.

[0105] 5. Differential protein screening

[0106] Differentially expressed proteins between the diseased and non-disease groups were screened based on a change in protein expression abundance greater than 1.2-fold and a T-test p-value less than 0.05. Figure 2As shown, compared with the population with lower degree of white matter lesions, the expression of 329 proteins was significantly different in the population with higher degree of white matter lesions. 190 proteins were upregulated in the high-lesion group and 139 proteins were downregulated in the high-lesion group.

[0107] 6. XGBoost and LASSO analysis

[0108] The importance scores and rankings of the differentially expressed proteins in different sample groups were calculated using the XGBoost machine learning model and the LASSO regression model. Forty-one of the most useful diagnostic protein features were then selected as plasma protein biomarkers that can be used to distinguish between diseased and non-diseased groups. Information on each protein biomarker is shown in Table 2.

[0109] Table 2 Biomarker Protein Information Table

[0110]

[0111] 7. Logistic Regression Analysis

[0112] The aforementioned protein biomarkers and clinical indicators were randomly combined, and a combined model was constructed using binary logistic regression. The optimal combination of protein biomarkers and clinical indicators was selected based on the AUC value. The optimal combination was four proteins and three clinical indicators (P05198 + P00966 + Q8IZP0 + P48061 + SDMT + MMSE + CTT_Trail1). A clinical diagnostic model was then constructed based on these protein biomarkers and clinical indicators. The parameters of the protein combined model and the protein-clinical indicator combined model are as follows: Figure 3 , 4 As shown in the figure. A nomogram is created based on the combined model to illustrate its application in diagnosis, such as... Figure 5-8 As shown.

[0113] The protein combinatorial model is shown in equation (1):

[0114] (1)

[0115] The model combining protein and clinical indicators is shown in formula (2):

[0116] (2)

[0117] Example 2: Testing and Evaluation of a White Matter Lesion Assessment Model

[0118] This embodiment uses 78 subjects from Zhongnan Hospital of Wuhan University as the test set, and uses a combination model of 4 proteins and 3 clinical indicators (formula (2) in embodiment 1) to verify the effect of assessing the degree of white matter lesions.

[0119] The clinical information of the subjects is shown in Table 3.

[0120] Table 3 Clinical information of subjects in the test set

[0121]

[0122] The sample source, pretreatment, LC-MS / MS detection, and data preprocessing in Example 2 are the same as in Example 1.

[0123] Test results are as follows Figure 9-11 As shown, AUC = 0.776 (0.671-0.881). The results indicate that the model has good discriminative power. This predictive model can effectively distinguish between high and low white matter lesion groups. The model's sensitivity reaches 65.7% and its specificity is 74.4%, showing that the model has good predictive ability and accuracy.

[0124] Example 3: Reagent Kit

[0125] This embodiment provides a kit comprising reagents for detecting protein biomarkers (P05198 + P00966 + Q8IZP0 + P48061) and clinical indicators (SDMT + MMSE + CTT_Trail1_time).

[0126] 3.1 Reagents for detecting protein biomarkers

[0127] Loading buffer 1 (10mM Tris-Cl, 1mM EDTA, 150mM KCl, 0.05% CHAPS);

[0128] Loading buffer 2 (10mM Tris-Cl, 1mM EDTA, 150mM KCl);

[0129] Lysis buffer (1% SDC / 100 mM Tris-HCl, pH=8.5 / 10 mM TCEP / 40 mM CAA);

[0130] trypsin;

[0131] SDB-RPS desalination column.

[0132] 3.2 Reagents for detecting clinical indicators

[0133] The reagents used to detect relevant clinical indicators.

[0134] Example 4: Computer Program

[0135] A program product related to leukoencephalopathy, the computer program product being used to perform a diagnostic procedure to determine the risk of a subject having leukoencephalopathy, includes the following steps:

[0136] (1) Obtain the expression level of each protein marker and the value of each clinical indicator in the plasma of the subject to be tested; the protein markers include P05198, P00966, Q8IZP0 and P48061; the clinical indicators include MMSE, CTTtrail1 and SDMT;

[0137] (2) Substitute the expression levels of each protein biomarker and the values ​​of clinical indicators into the binary logistic regression equation to calculate the logarithm y of the dominance of the subject to be tested;

[0138] The formula for the binary logistic regression equation is as follows:

[0139]

[0140] Where A is the intercept term, B1 to B6 are the regression coefficients of the independent variables; x1, x2, x3, and x4 are the expression levels of proteins P05198, P00966, Q8IZP0, and P48061, respectively; and x5, x6, and x7 are the expression levels of clinical indicators SDMT, MMSE, and CTTtrail1, respectively.

[0141] A=1.51, B1=0.41, B2=-0.3, B3=-0.22, B4=0.12, B5=-0.08, B6=-0.04, B7= 0.02.

[0142] (3) Calculate the probability P that the subject is a healthy person based on y, P=exp(y) / {1+exp(y)}, where exp(y) represents the natural exponential function;

[0143] (4) Based on the comparison between probability P and reference value, diagnose or predict whether the subject has leukoencephalopathy or is at risk of having leukoencephalopathy.

[0144] Understandably, the probability P is generally set to 0.5, but it can be adjusted as needed based on the actual situation.

[0145] The above description is merely a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications, equivalent substitutions, and improvements made by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the invention.

Claims

1. A combination of protein biomarkers, characterized in that, include: P05198, P00966, Q8IZP0 and P48061.

2. The application of a reagent for detecting a combination of protein biomarkers in the preparation of products for diagnosing white matter lesions, wherein the combination of protein biomarkers includes P05198, P00966, Q8IZP0 and P48061.

3. A combination of a protein biomarker and a clinical indicator, characterized in that, The protein biomarkers include: P05198, P00966, Q8IZP0, and P48061; the clinical indicators include the Mini-Mental State Examination (MMSE), the Color-Linked Number Test (CTTTtrail1), and the Symbolic Number Transformation Test (SDMT).

4. The application of a combination of a protein biomarker detection reagent and a clinical indicator detection product in the preparation of a diagnostic product for leukoencephalopathy, wherein the protein biomarker combination includes P05198, P00966, Q8IZP0, and P48061; and the clinical indicator includes MMSE, CTTtrail1, and SDMT.

5. A reagent kit, characterized in that, The kit includes a detection reagent for detecting a combination of the protein biomarkers and clinical indicators as described in claim 3.

6. The use of the kit of claim 5 in the preparation of a product for detecting leukoencephalopathy, and the use of the detection reagent in the kit of claim 5 in the preparation of a kit for diagnosing leukoencephalopathy.

7. A computer program product related to white matter lesions, characterized in that, The computer program product is used to diagnose the risk of a subject having leukoencephalopathy, including the following steps: The expression levels of each protein biomarker and the values ​​of each clinical indicator in the plasma of the subjects to be tested were obtained; the protein biomarkers included P05198, P00966, Q8IZP0 and P48061; the clinical indicators included MMSE, CTTtrail1 and SDMT. Substitute the expression levels of each protein biomarker and the values ​​of clinical indicators into the binary logistic regression equation to calculate the logarithm y of the advantage of the subject. Calculate the probability P that the subject is a healthy person based on y, P = exp(y) / {1 + exp(y)}, where exp(y) represents the natural exponential function; Based on the comparison between probability P and reference value, the test subject is diagnosed or predicted to have leukoencephalopathy or to be at risk of having leukoencephalopathy.

8. The program product according to claim 7, characterized in that, The formula for the binary logistic regression equation is: Where A is the intercept term, B1 to B7 are the regression coefficients of the independent variables; x1, x2, x3, and x4 are the expression levels of proteins P05198, P00966, Q8IZP0, and P48061, respectively; and x5, x6, and x7 are the expression levels of clinical indicators SDMT, MMSE, and CTTtrail1, respectively.

9. The program product according to claim 7 or 8, characterized in that, The method for obtaining the binary logistic regression equation includes: The expression levels of protein markers and the values ​​of clinical indicators were obtained in healthy controls and patients with white matter lesions. Using the expression levels of protein biomarkers and the values ​​of clinical indicators as inputs, and whether or not a disease is present as output, a binary logistic regression equation is obtained through training.

Images