Plasma proteins as biomarkers for diagnosis of STEMI with hypertension and application

Abstract

The application discloses a plasma protein as a biomarker for diagnosing STEMI combined with hypertension and application, and has the characteristics that the plasma protein biomarker is CA1, and can be applied to preparation of a product for diagnosing STEMI combined with hypertension; and the advantages are that early diagnosis of STEMI combined with hypertension is realized through determination of the plasma protein biomarker, an effective tool is provided for diagnosis of STEMI combined with hypertension, and the detection efficiency is high and the pertinence is strong.

Description

Technical Field

[0001] This invention relates to diagnostic techniques for STEMI complicated with hypertension, and in particular to a plasma protein as a biomarker for diagnosing STEMI complicated with hypertension and its application. Background Technology

[0002] Acute ST-segment elevation myocardial infarction (STEMI) is a severe emergency caused by acute coronary artery occlusion leading to myocardial ischemia and necrosis. It is characterized by rapid onset, rapid progression, and high mortality and disability rates. Although the widespread adoption of reperfusion strategies such as percutaneous coronary intervention (PCI) has significantly improved overall prognosis, in real-world clinical practice, STEMI patients still face a high risk of in-hospital death, reperfusion injury, severe arrhythmias, cardiogenic shock, and multiple organ dysfunction. Therefore, rapid and accurate risk stratification and prognostic assessment of STEMI patients in the early stages of the disease, along with precise monitoring and intervention for high-risk groups, are crucial clinical needs in the field of cardiovascular emergencies and critical care.

[0003] It is noteworthy that hypertension, as one of the most common cardiovascular risk factors, has an extremely high comorbidity rate in STEMI patients. Previous studies have shown that approximately 50%–70% of STEMI patients have a history of hypertension to varying degrees. Hypertension is not only a significant driver of the formation and progression of coronary atherosclerosis, but also a key regulatory factor influencing the pathophysiological state, clinical manifestations, and prognosis of STEMI patients. On the one hand, long-term hypertension can significantly reduce the myocardium's tolerance to acute ischemic events by inducing endothelial dysfunction, vascular remodeling, left ventricular hypertrophy, and microcirculatory damage. On the other hand, excessive activation of the sympathetic nervous system and the renin-angiotensin-aldosterone system (RAAS) during acute STEMI attacks often leads to drastic fluctuations or even loss of blood pressure control, further exacerbating myocardial damage, heart failure, and the risk of complications. Numerous clinical studies have shown that STEMI patients with hypertension often have more complex pathological bases, more significant individual heterogeneity, and worse short-term and long-term prognoses. However, current strategies for early identification, refined risk assessment, and individualized intervention for the high-risk subgroup of "STEMI combined with hypertension" are still significantly insufficient.

[0004] Current clinical assessments mainly rely on myocardial injury markers (such as troponin and CK-MB), blood pressure levels, and some comprehensive risk scoring systems (such as the GRACE score). These methods have some value in general risk stratification, but they still have many limitations in the clinical context of STEMI with hypertension: (1) Traditional myocardial injury indicators mainly reflect the degree of myocardial necrosis and are difficult to fully reflect the background state of hypertension-related chronic vascular injury, endothelial dysfunction, and inflammatory activation; (2) Single or short-term blood pressure measurements are greatly affected by acute stress, drug intervention, and monitoring time points, making it difficult to accurately characterize the long-term impact of hypertension on the course and prognosis of STEMI; (3) There is currently a lack of specific biomarkers or multi-protein combined biomarker combinations for this subgroup, making it difficult to achieve rapid, objective, and widely applicable early stratified diagnosis; (4) Current detection strategies are mostly simple superpositions of single indicators or a few clinical parameters, and have not fully utilized the complex molecular network information revealed by proteomics, thus limiting the practical application of precision medicine strategies in STEMI patients with hypertension.

[0005] During STEMI, acute coronary artery occlusion can rapidly trigger myocardial ischemia and necrosis, accompanied by a strong inflammatory response, platelet activation, and coagulation cascade. In STEMI patients with hypertension, long-term endothelial damage, increased arterial stiffness, and microvascular dysfunction amplify these pathological processes, further accompanied by RAAS activation, enhanced oxidative stress, and an imbalance in the immune-inflammatory response. These pathological changes can lead to significant alterations in the plasma protein profile. For example, inflammation-related proteins (such as interleukins and chemokines), proteins related to endothelial function and vascular remodeling (such as adhesion molecules and matrix metalloproteinases), coagulation and fibrinolysis-related proteins, and myocardial stress and injury-related proteins exhibit dynamic abnormalities in plasma. Therefore, plasma proteins not only indirectly reflect the degree of myocardial injury but also provide a more comprehensive characterization of the systemic inflammatory state, vascular dysfunction, and thrombotic risk in STEMI patients with hypertension.

[0006] With the development of proteomics technology, especially the application of high-throughput detection platforms based on plasma / serum or exosome-derived proteins (such as Olink PEA), it has become possible to discover differentially expressed proteins related to diseases at a global level, screen for highly specific biomarkers, and construct diagnostic models. Compared with traditional biochemical indicators, protein biomarkers can more directly reflect key pathological processes such as acute-phase inflammatory immune activation, endothelial dysfunction, coagulation and fibrinolytic system disorders, metabolic stress, and organ interactions. Therefore, developing protein biomarkers for the early identification and risk prediction of STEMI complicated with hypertension, especially candidate proteins with clear biological significance that can be translated into clinical diagnostic kits, has significant clinical value and promising industrial application prospects. Summary of the Invention

[0007] The technical problem to be solved by this invention is to provide a plasma protein as a biomarker for diagnosing STEMI complicated with hypertension and its application. Early diagnosis of STEMI complicated with hypertension through the measurement of plasma protein biomarkers provides an effective tool for the diagnosis of STEMI complicated with hypertension, with high detection efficiency and strong specificity.

[0008] The technical solution adopted by the present invention to solve the above-mentioned technical problems is as follows: a plasma protein is used as a biomarker for diagnosing STEMI complicated with hypertension, wherein the plasma protein biomarker is CA1.

[0009] Compared with the prior art, the advantages of the present invention are as follows: I. This invention detects STEMI complicated with hypertension in its early stages, even asymptomatically, by detecting the specific protein marker CA1 in plasma. This molecular-level detection method can identify early biological changes associated with the lesion, without relying on structural changes in blood vessels, thus enabling diagnosis at a much earlier stage.

[0010] II. Plasma testing can be completed through simple blood sample collection and is a non-invasive test. Liquid biopsy is not only convenient but also allows for multiple samplings for follow-up and monitoring of disease progression, making it particularly suitable for long-term management and regular screening. The specific protein marker CA1 in plasma can more accurately reflect the pathophysiological processes (such as inflammation and vascular wall degradation) associated with STEMI and hypertension.

[0011] Third, this invention assesses each patient's pathological condition and provides personalized diagnostic and treatment plans by detecting the plasma protein biomarker profile containing CA1. This molecular marker-based personalized assessment helps guide early intervention and optimize treatment decisions.

[0012] Fourth, this invention, through the detection of the plasma protein biomarker CA1, can expand the scope of screening, making it applicable not only to high-risk groups but also to other potentially at-risk populations for early screening. Plasma testing provides a broader and more accurate screening tool. This test is low-cost and simple to operate, making it suitable for screening in primary healthcare units or resource-limited areas. This liquid biopsy can serve as an effective early diagnostic tool, reducing reliance on high-cost imaging equipment and optimizing the allocation of medical resources.

[0013] Fifth, this invention can significantly improve the early detection capability and long-term management effect of STEMI combined with hypertension, and fill the gaps in the existing diagnostic technology in early detection and personalized assessment.

[0014] Furthermore, the CA1 is derived from human peripheral blood.

[0015] Furthermore, the present invention also provides the application of the plasma protein biomarker in the preparation of a product for diagnosing STEMI complicated with hypertension.

[0016] Furthermore, the product is a reagent kit.

[0017] Furthermore, the kit includes reagents or biochips for detecting CA1 expression levels.

[0018] Furthermore, the biochip includes a microarray biochip. Attached Figure Description

[0019] Figure 1 The figures show the distribution and analysis of plasma proteins. A is a heatmap of all expressed plasma proteins between the STEMI group with hypertension and the control group, where STEMI represents the STEMI group with hypertension and Control represents the normal control group. B is a NPX distribution diagram of all expressed plasma proteins between the STEMI group with hypertension and the control group, where QC_Warning samples are marked in red (Warning) and Pass samples are marked in cyan (Pass). C is a principal component analysis plot of all expressed plasma proteins. Figure 2 The data is presented as a plasma protein profile. A is a volcano plot of plasma proteins, where red dots represent upregulated proteins, gray dots represent insignificant proteins, and blue dots represent downregulated proteins. B shows the number of differentially expressed proteins (DEPs) in the two modules. C is a heatmap of plasma proteins, displaying the expression profiles of five proteins at absolute logFC > 0.5 and p < 0.05 cutoff values, where STEMI with hypertension represents the STEMI with hypertension group, and Control represents the normal control group. D is a dimensionality reduction analysis plot of the plasma protein expression profiles of STEMI with hypertension and control samples. E is a Pearson correlation analysis plot of normal samples and STEMI with hypertension samples. Figure 3Core protein maps were selected for machine learning. A represents the LASSO regression cross-validation curve, showing the trend of binomial bias with log(λ), with the dashed line indicating the optimal penalty parameter. B shows the trajectory of regression coefficients for each candidate protein under different λ values, with the dashed line marking the optimal λ. C shows the importance ranking of the final model protein features. D compares the performance of various machine learning models in terms of AUC, accuracy (ACC), and F-beta, sorted by AUC, with color intensity representing the magnitude of the indicator. E shows the ROC curve and corresponding AUC value of the optimal model. F shows the calibration curve constructed based on the bootstrap method. G shows the confusion matrix of the model classification results in the test set. Figure 4 A diagram illustrating the biological significance of target genes for plasma core proteins; a PPI network among potential biomarkers was constructed using the STRING database, with edge width representing the overall score; PPI stands for protein-protein interaction; Figure 5 The graphs show the expression of core proteins in the ELISA-validated external cohort of the First Affiliated Hospital of Ningbo University and the extrapolation of core proteins carried by plasma proteins to independent external cohorts. In the graphs, A represents the expression of core proteins in the external cohort as validated by ELISA; B represents the ROC curve of the core proteins in the external cohort as validated by ELISA; C represents the calibration curve of the core proteins in the external cohort as validated by ELISA; and D represents the confusion matrix of the core proteins in the external cohort as validated by ELISA. Detailed Implementation

[0020] The present invention will be further described in detail below with reference to the accompanying drawings and embodiments. However, the embodiments of the present invention are not limited thereto. Unless otherwise specified, the reagents, methods, and equipment used in the present invention are conventional reagents, methods, and equipment in this technical field. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those familiar with the art. Furthermore, any methods and materials similar to or equivalent to those described herein may be applied to the present invention.

[0021] Example 1: Experimental Materials This invention collected 17 STEMI patients with hypertension and 11 healthy controls from the First Affiliated Hospital of Ningbo University as an internal cohort, and used PEA technology based on the Olink platform to analyze the proteomic expression of circulating plasma in STEMI patients with hypertension.

[0022] The inclusion criteria for STEMI patients with hypertension include: (1) meeting the diagnostic criteria of the latest version of the "Guidelines for the Diagnosis and Treatment of Acute ST-Segment Elevation Myocardial Infarction"; (2) meeting the diagnostic criteria of the "Guidelines for the Prevention and Treatment of Hypertension in China"; (3) the time from the onset of STEMI to admission is ≤24 hours (or the time window is limited according to the actual design); (4) signing an informed consent form and agreeing to the storage and analysis of blood samples.

[0023] Exclusion criteria for STEMI with hypertension include: (1) non-ST-segment elevation myocardial infarction or unstable angina; (2) secondary hypertension; (3) severe complications, including severe liver failure, severe renal failure, active infection, or malignant tumor; (4) a history of major cardiovascular events, including myocardial infarction within the past 6 months, previous heart transplantation, or severe structural heart disease; and (5) pregnancy or lactation.

[0024] This invention recruited patients diagnosed with STEMI and hypertension, as well as healthy volunteers. Venous blood was collected using whole blood collection EDTA tubes (BD vacuum blood collection tubes) to inhibit clotting. Blood samples were centrifuged at 1500×g for 20 minutes at 4°C to remove cells. The supernatant was collected and centrifuged at 3000×g for 15 minutes at 4°C. The supernatant (plasma) was then collected, and the first 2 mL of plasma from each sample was rapidly frozen at -80°C to maintain the stability and integrity of plasma biomolecules.

[0025] Example 2: Experimental Method 1. Plasma proteins were extracted from plasma samples using the ExoQuick Plasma Protein Kit (EXOQ5A1; SystemBiosciences, LLC, USA), strictly adhering to the manufacturer's protocol. The specific steps were as follows: Mix 250 μL of plasma sample with 36 μL of rapid plasma precipitation solution from the ExoQuick kit and incubate at 4 °C for 30 min. Then centrifuge at 1500 × g for 30 min. Discard the supernatant and centrifuge again at 1500 × g for 5 min. Then resuspend the obtained plasma nanoparticles (plasma) in 100 μL of sterile phosphate-buffered saline (PBS) and store at -80 °C for subsequent analysis.

[0026] 2. Differentially expressed plasma proteins were obtained using Olink proteomics technology, cluster analysis, and biological function enrichment analysis. The specific steps are as follows: Olink proteomics technology is an innovative protein detection and quantification technology that combines protein and antibody immunoreaction technology with oligonucleotide amplification technology based on Proximity Extension Assay (PEA).

[0027] Olink protein biomarker assays are divided into the Target series (48-fold and 96-fold) and the Explore series (384-fold and 3072-fold). Target proteins are detected using qPCR. Specifically, the prepared protein samples are diluted according to the protein concentration of different target monitoring panels, then co-incubated with antibodies linked to oligonucleotide sequences. After base pairing of the two oligonucleotide sequences linked to the protein, the paired oligonucleotide sequences are extended in the reaction system under the action of DNA polymerase. Pre-amplification is followed by PCR detection. The Olink Target 96 series can be used to analyze targeted protein biomarkers of specific proteins associated with specific diseases or biological functions (Ding Z, Wang N, Ji N, et al. Proteomics technologies for cancer liquid biopsies[J]. Molecular Cancer, 2022, 21(1):1-11). Compared with previous protein immunoassay techniques, this technology can provide new insights into disease processes, improve disease detection, and contribute to a better understanding of biology.

[0028] 2.1 Total protein extraction and immunolinking Total plasma protein extraction: Total protein was extracted from the plasma PBS suspension obtained in step 1, ensuring that the plasma sample remained stable when stored at -80°C; Plasma protein extraction reagent: RIPA lysis buffer (Thermo Scientific, #89900), with 1× phosphatase inhibitor and PMSF added; Protein concentration determination: BCA protein quantification kit (Thermo Scientific, #23225).

[0029] Extraction steps: Add an equal volume of RIPA lysis buffer to the plasma suspension, mix thoroughly, and incubate in an ice bath for 30 minutes; centrifuge at 13,000×g for 10 minutes at 4°C, and collect the supernatant; determine the protein concentration using the BCA method, dilute and store for subsequent detection.

[0030] Immune linking of oligonucleotide sequence antibodies: Antibody and sequence design: The oligonucleotide sequence antibody was purchased from Olink Proteomics.

[0031] Immunolinking procedure: Mix the protein sample with the oligonucleotide sequence antibody at the recommended concentration and incubate on ice for 30 minutes; Base complementation ligation: Two oligonucleotide sequences were ligated to different antibodies targeting the protein. After complementation ligation, DNA polymerase (Thermo Scientific, #EP0502) was added to extend the reaction. Reaction system: reaction buffer (10 mM Tris-HCl, pH 8.0; 50 mM KCl; 5 mM MgCl2). Base complementation reaction conditions: incubation at 25°C for 30 minutes. DNA polymerization reaction conditions: reaction at 37°C for 20 minutes, followed by heating at 95°C for 5 minutes to terminate the reaction.

[0032] 2.2 Sample library preparation and qPCR detection In the library preparation stage, a specific barcode tag is first added to each sample after the extension reaction for subsequent library differentiation and sample identification. The equipment required for library preparation includes an Applied Biosystems QuantStudio 5 qPCR instrument, as well as barcode primers and library preparation reagents purchased from Olink Proteomics. In practice, the primers with barcode tags are thoroughly mixed with the extension reaction product, followed by a second amplification to improve specificity and signal intensity. The amplified product is then purified using magnetic beads (e.g., using AMPure XP magnetic beads, Beckman Coulter) to obtain high-purity library samples, providing high-quality templates for subsequent qPCR detection.

[0033] In the qPCR detection stage, the purified library sample was added to the qPCR reaction system at an appropriate dilution factor. The total reaction volume was 20 μL, including 10 μL of 2× qPCR Master Mix (e.g., Thermo Fisher, #A25742), 1 μL of library sample template, 0.4 μL of target primer pair (final concentration 200 nM), and 8.6 μL of sterile ultrapure water. In the qPCR reaction, initial denaturation was performed at 95°C for 5 minutes to ensure complete denaturation of the DNA template; this was followed by 40 cycles, including heating at 95°C for 15 seconds and annealing extension at 60°C for 1 minute, to complete library amplification.

[0034] After amplification, the fluorescence signal was acquired in real time using the software (QuantStudio Design & Analysis Software) included with the qPCR instrument, and the Ct value of each sample was recorded. The Ct value was then converted into NPX value (Olink Normalized Protein eXpression) using data analysis software to provide standardized protein expression data for subsequent differential expression analysis.

[0035] 2.3 Convert the .bml or .bcl files generated by qPCR detection into Ct values ​​or counts files to obtain the data input files required by the NPX Manager software. Then, preprocess the qPCR quantitative data using NPX Signature software, and import it into Olink NPX Signature (NPX Manager) software for sample type annotation, protein NPX calculation, data quality control, data export, and saving of the Certificate of Analysis (CoA). This yields the input files required for the bioinformatics analysis workflow: NPX data. Box plots of the original sample NPX data distribution and principal component analysis (PCA) plots of the samples are also generated.

[0036] 2.4 Based on the original sample NPX data distribution box plot and sample principal component analysis (PCA) plot, check the sample dimensionality reduction clustering situation to see if there are outliers, so that subsequent analysis can decide whether to remove outliers; then, create a heatmap of all protein expression clusters to gain a preliminary understanding of the overall protein expression pattern in this application, and also to identify proteins with abnormal expression.

[0037] 2.5 After the raw data quality control above, clean data of protein NPX expression will be obtained. Then, protein expression clustering heatmap analysis, protein correlation coefficient heatmap analysis, and expression box plot analysis can be performed on each monitoring panel. Based on clean data of protein NPX expression, this invention performs Student's test or analysis of variance on different experimental variables to obtain differentially expressed proteins under different experimental conditions. Visual analysis of the differentially expressed protein data is then performed, including drawing differentially expressed protein volcano plots and monitoring panel differentially expressed protein clustering heatmaps. This invention also performs Pearson correlation analysis and UMAP analysis to assess the similarities and differences in plasma protein expression between STEMI patients with hypertension and healthy controls.

[0038] 3. The machine learning modeling and feature selection method of this invention selects the optimal protein. The machine learning modeling and feature selection method employs preprocessing to ensure data quality and comparability. Differentially expressed proteins obtained through screening are used as candidate features to construct a feature matrix with protein expression level as the independent variable and STEMI combined with hypertension as the dependent variable.

[0039] In the feature selection stage, this invention uses the feature importance scoring method based on the random forest algorithm to rank and screen candidate protein features by importance, and identify the key protein molecules that contribute the most to the model prediction.

[0040] In the model building phase, various mainstream machine learning algorithms were employed to establish predictive models, including Logistic Regression, Support Vector Machine (SVM), Random Forest, Extreme Gradient Boosting Tree (XGBoost), K-Nearest Neighbor (KNN), and Artificial Neural Network (ANN). A systematic comparative analysis of these different models was conducted. K-fold cross-validation was used to train and validate each model, and repeated sampling and cross-validation reduced the risk of overfitting.

[0041] To evaluate the robustness and generalization ability of the model, bootstrap resampling and repeated cross-validation were used for internal validation. AUC, calibration curve, Brier score, and decision curve analysis (DCA) metrics were calculated during internal validation. Simultaneously, ELISA test data from independent external validation cohorts were input into the optimal machine learning model for predictive analysis, providing a systematic review of the model's diagnostic efficacy and clinical applicability across different sample sources.

[0042] 4. This invention also performs biological function enrichment analysis on the differentially expressed protein data. All identified plasma proteins associated with STEMI and hypertension were annotated using the STRING database, and protein-protein interaction (PPI) analysis was performed.

[0043] 5. The expression level of the core protein CA1 in the external cohort was further verified using enzyme-linked immunosorbent assay (ELISA). The specific steps are as follows: Peripheral blood samples were collected via EDTA anticoagulant tubes and centrifuged at 3,000 × g for 10 minutes at 4°C to separate plasma. The supernatant was aliquoted and stored at −80°C for later use. Before testing, plasma samples were slowly thawed at 4°C and gently mixed.

[0044] Following the instructions of each ELISA kit, the plasma sample was diluted at the specified ratio and added to a 96-well plate coated with antibody, and incubated at 37°C. After washing, biotin-labeled antibody and horseradish peroxidase-labeled streptavidin were added sequentially. After the colorimetric reaction was completed and the reaction was terminated, the absorbance was measured at 450 nm using an ELISA reader.

[0045] The concentrations of each target protein were calculated based on the standard curve to assess the differences in their expression levels among different inventor populations.

[0046] 6. Analyze the aforementioned experimental results. Specifically, all statistical tests in the aforementioned experimental methods were two-tailed tests; p-values ​​less than 0.05 and false detection rates (FDRs) less than 0.05 were considered statistically significant; descriptive statistics for normally distributed continuous variables were reported as mean ± standard deviation; comparisons of continuous variables were performed using the corresponding Wilcoxon rank-sum test or Student's t-test; categorical variables were analyzed using the chi-square test or Fisher exact test; all data processing, statistical analysis, and graphing were performed using R software (version 4.4.3).

[0047] Example 3: Experimental Results I. Detection of plasma protein biomarkers in STEMI patients with hypertension using Olink proteomics technology This invention analyzed 92 protein biomarkers in the plasma of STEMI patients with hypertension and healthy controls, and the results are as follows: Figure 2 As shown in the figure. The distribution and analysis of plasma proteins are shown in the figure below. Figure 1 As shown. A heatmap of all expressed plasma proteins between the STEMI combined hypertension group and the control group ( Figure 1 (See Figure A). The distribution patterns of plasma proteins in different samples can be visually observed. The results show significant differences in protein expression between the STEMI with hypertension group and the healthy control group, with some proteins exhibiting unique expression patterns in both groups. Further NPX distribution box plot analysis of all expressed plasma proteins between the STEMI with hypertension group and the control group (…) Figure 1 (Figure B) shows that the NPX (normalized protein expression) levels between the STEMI group with hypertension and the healthy control group exhibited significant differences across multiple proteins, revealing the protein expression characteristics between the two groups. Principal component analysis (PCA) plots of all expressed plasma proteins (Figure B) Figure 1 The results (C) showed that the samples were clearly separated in the principal component space, indicating that the differences in plasma protein expression characteristics between STEMI patients with hypertension and healthy controls were significant and consistent. These results collectively support the possibility that plasma proteins can serve as potential biomarkers to differentiate STEMI patients with hypertension from healthy individuals, thus enabling the diagnosis of STEMI with hypertension.

[0048] The significance criterion used in the analysis was P < 0.05, based on the t-test or ANOVA results of the differential expression analysis. The calculation of this threshold further validated the statistical significance of plasma protein expression levels between the two groups. Validation was performed using an internal cohort. Figure 2(A, B, C) It was found that there were significant differences in the expression of CA1 protein biomarkers in the plasma of STEMI patients with hypertension and healthy controls. Among them, the expression of 70 proteins was significantly upregulated and the expression of 2 proteins was significantly downregulated in the plasma of STEMI patients with hypertension.

[0049] Regarding the cohort setup, the purpose of this invention is to ensure the stability of the screened plasma protein biomarkers. The internal cohort is used for the preliminary discovery of potential differentially expressed proteins, and this invention ensures that the screened plasma proteins provide solid experimental evidence for their potential diagnostic markers for STEMI complicated with hypertension.

[0050] In addition, this invention also performed Pearson correlation analysis and UMAP analysis to assess the differences in plasma protein expression between STEMI patients with hypertension and healthy controls. Figure 2 (D, E). These results show that there are significant differences in plasma protein expression between STEMI patients with hypertension and healthy individuals.

[0051] II. Screening of target proteins in plasma proteome for STEMI patients with hypertension This invention uses machine learning technology to screen for the target protein, CA1, based on differentially expressed protein data. The method for screening plasma protein biomarkers includes the following steps: Step 1: Using a plasma protein extraction kit, plasma proteins were extracted from the peripheral blood samples of the subjects, and quality control and quantitative detection were performed on the peripheral blood samples; the main operation process is as follows: Step 1.1. Total protein extraction and immunolinking: Total protein was extracted from the obtained plasma PBS suspension, and the protein concentration was determined using a BCA protein quantification kit. Immunolinking and base complementation ligation of oligonucleotide sequence antibodies: Immunolinking steps: Mix protein sample with oligonucleotide sequence antibody and incubate; Base complementation ligation: Two oligonucleotide sequences were ligated to different antibodies targeting the protein; after complementary ligation, DNA polymerase was added to extend the system. The reaction system consisted of: reaction buffer (10 mM Tris-HCl, pH 8.0; 50 mM KCl; 5 mM MgCl2); base complementation reaction conditions: incubation at 25°C for 30 minutes; DNA polymerization reaction conditions: reaction at 37°C for 20 minutes, followed by heating at 95°C for 5 minutes to terminate the reaction.

[0052] Step 1.2. In the library preparation stage, after the extension reaction, a specific barcode tag is first added to each sample for subsequent library differentiation and sample identification, and then purified to obtain the library sample. In the qPCR detection stage, the purified library sample is diluted and added to the qPCR reaction system to complete the amplification of the library sample. After amplification, the fluorescence signal is acquired in real time using a qPCR instrument, and the Ct value of each sample is recorded. The Ct value is converted into NPX value using data analysis software to provide standardized protein expression data for subsequent differential expression analysis. The total reaction volume was 20 μL, including 10 μL of 2× qPCR Master Mix, 1 μL of library template, 0.4 μL of target primer pair (final concentration 200 nM), and 8.6 μL of sterile ultrapure water. In the qPCR reaction, initial denaturation was performed at 95°C for 5 minutes to ensure complete denaturation of the DNA template; this was followed by 40 cycles, including heating at 95°C for 15 seconds and annealing extension at 60°C for 1 minute.

[0053] Step 1.3. Convert the data generated by qPCR detection into Ct values ​​or counts files, process them with NPXSignature software and Olink NPX Signature (NPX Manager) software to obtain the original sample NPX data required for the bioinformatics analysis workflow, and create box plots of the original sample NPX data distribution and principal component analysis (PCA) plots of the samples. Step 1.4. Based on the box plot of the original sample NPX data distribution and the principal component analysis (PCA) plot of the sample, check the dimensionality reduction and clustering of the sample; after quality control of the original data, clean data of protein NPX expression are obtained, and differential proteins under different experimental variables are obtained by performing Student's test or analysis of variance on different experimental variables. Step 1.5. Perform biological function enrichment analysis on the differentially expressed protein data, including GO biological processes, GO molecular functions, KEGG pathways, human gene sets, Reactome gene sets, and typical pathways.

[0054] Step 2: Based on the differentially expressed proteins obtained in Step 1, a candidate protein set is established. Step 3: Based on the candidate protein set obtained from the aforementioned differentially expressed protein analysis among different groups, LASSO (Least Absolute Shrinkage and Selection Operator) logistic regression is used to regularly screen the candidate protein set. The optimal penalty parameter λ is determined through cross-validation to obtain the protein set corresponding to non-zero regression coefficients, which serves as the initial screening key protein feature set with stable predictive contributions. Based on the LASSO initial screening results (i.e., the initial screening key protein feature set), a random forest model is further constructed and feature importance is calculated. The initial screening proteins in the initial screening key protein feature set are ranked by contribution. The initial screening protein with the highest contribution is selected according to the importance ranking to form the optimal protein combination feature set for subsequent modeling. Various machine learning algorithms are used to establish classification models (such as logistic regression, support vector machine, K-nearest neighbors, Naive Bayes, linear discriminant analysis, random forest, gradient boosting / XGBoost). The algorithm is trained and optimized on the training set using a leave-out validation and unified hyperparameter search strategy. By comparing the comprehensive performance of each algorithm on the validation set (including AUC, accuracy, F1, etc.), the algorithm with the best comprehensive performance and the best stability is selected as the core algorithm of the diagnostic model of this invention. Step 4: Based on the key proteins obtained through screening, a protein-protein interaction (PPI) network is constructed using a protein interaction database. Through topological structure analysis, core node identification, and signal pathway mapping, the molecular regulatory mechanism of key proteins in the occurrence and development of STEMI complicated with hypertension is analyzed. Step 5: Input the external validation data of ELISA into the optimal machine learning model for predictive analysis, and evaluate the diagnostic performance and clinical applicability of the model in independent samples.

[0055] This invention employs LASSO regression to screen candidate features and determines the optimal regularization parameter through cross-validation, thereby obtaining stable combinations of protein features. The relevant screening results are illustrated in the figure below. Figure 3 As shown. The LASSO regression cross-validation curve is as follows. Figure 3 As shown in Figure A, the trajectory of the regression coefficients of each candidate protein under different λ values ​​is as follows. Figure 3 As shown in Figure B, this invention constructs various machine learning classification models and systematically evaluates their predictive performance. Figure 3 The results showed that proteins such as CA1, AK1, PDAP1, MYL4, POMC, GAS6, and TCTN3 made significant contributions to the model. In the multi-model comparative analysis, different algorithms showed differences in AUC, accuracy, and F-beta value. The optimal model demonstrated high discriminative ability in distinguishing STEMI patients from healthy controls. Figure 3As shown in Figure D. ROC curve analysis further confirms that the model has good classification performance, such as... Figure 3 As shown in E. Figure 3 The F-square is used to evaluate the consistency between the model's predicted probabilities and the actual outcomes. For example... Figure 3 As shown in Figure F, the calibration curve results show that the model's predicted probability has good consistency with the actual outcome, indicating that it has high stability and reliability. Figure 3 The G-analysis showed that the model could accurately distinguish between disease group and control group samples in the test set. Based on the above results, the CA1-based predictive model constructed in this invention has potential clinical application value in the early identification and risk assessment of STEMI.

[0056] This invention enables early diagnosis of STEMI patients with hypertension by combining the detection of multiple protein biomarkers and establishing mathematical models or algorithms (such as machine learning, statistical models, etc.).

[0057] III. Comprehensive functional enrichment of target proteins in the plasma proteome of STEMI patients with hypertension In the aforementioned experimental method of this invention, a PPI interaction network labeled with UniProt ID and gene symbols was generated through protein-protein interaction (PPI) analysis. After removing isolated target genes lacking interaction and applying a confidence threshold of 900, the gene interaction network was visualized using Cytoscape (as shown in Figure 4).

[0058] IV. ELISA Validation of Differentially Expressed Proteins in Circulating Plasma of STEMI Patients with Hypertension In circulating plasma from STEMI patients with hypertension and healthy controls, the differentially expressed protein with the highest predictive performance was identified as the pivot protein CA1. The criteria for selecting CA1 as a pivot protein were primarily based on the following quantitative indicators: (1) meeting the significance threshold (P<0.05) in differential expression analysis; (2) exhibiting high predictive ability in machine learning models; and (3) possessing high connectivity or a key position in protein-protein interaction (PPI) networks. Based on these quantitative criteria, the key roles of these proteins were further verified experimentally.

[0059] This invention used ELISA to validate the expression of the pivot protein CA1 in an external validation cohort (12 patients and 8 healthy controls). ELISA results showed that the circulating plasma CA1 levels in STEMI patients with hypertension were significantly higher than those in the healthy control group (P<0.05). Furthermore, Figure 5The ROC curve of B further evaluated the potential of these proteins as diagnostic biomarkers, with an AUC value of 0.823 (0.593-1.00), showing extremely strong diagnostic ability and significantly distinguishing STEMI patients with hypertension from healthy individuals.

Claims

1. The application of a reagent for detecting plasma protein biomarkers in the preparation of products for diagnosing STEMI complicated with hypertension, characterized in that, The plasma protein biomarker is CA1.

2. The application according to claim 1, characterized in that, The product in question is a reagent kit.

3. The application according to claim 2, characterized in that, The kit includes reagents or biochips for detecting CA1 expression levels.

4. The application according to claim 3, characterized in that, The biochip includes a microarray biochip.

Images