An early screening marker related to immune response in ischemic stroke based on Olink proteomics and its selection method and application
Through Olink proteomics technology and machine learning methods, six immune response-related protein markers were identified, and a high sensitivity and specificity ischemic stroke early screening model was established, which solved the sensitivity and sensitivity of the early screening model in the existing technology. The problem of insufficient specificity has achieved high accuracy detection and pathological mechanism interpretation of ischemic stroke.
Patent Information
- Application Number
- CN202410062248.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-01-16
- Publication Date
- 2024-11-22
- Estimated Expiration
- 2044-01-16
AI Technical Summary
The sensitivity and specificity of the existing ischemic stroke early screening model is insufficient, and traditional methods cannot detect immune response-related proteins with high sensitivity and specificity, affecting the accuracy of early diagnosis.
Using Olink proteomics-based technology, six protein markers related to immune responses (MASP1, STC1, HCLS1, CLEC4D, PTH1R and PIK3AP1) were identified through Olink proteomic detection and machine learning methods to establish high sensitivity and specificity. early screening model for assisted diagnosis of ischemic stroke.
High accuracy detection of early screening of ischemic stroke is achieved, providing a biological explanation of the pathological process of immune response, and improving the efficiency of early screening and diagnosis accuracy.
Smart Images

Figure CN117969851B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of biomedical technology, and in particular to an ischemic stroke immune response-related early screening marker based on Olink proteomics and an application thereof. Background Art
[0002] Stroke is one of the main causes of disability and death in humans, and most of these patients suffer from ischemic stroke. Rapid and accurate screening of ischemic stroke in the acute phase is crucial to gaining the treatment time window and reducing mortality, disability, and recurrence risks. Currently, the main basis for clinical screening comes from doctors' judgment and imaging examinations, but diseases with similar clinical manifestations to ischemic stroke may interfere with doctors' judgments, and imaging examinations have problems of insufficient sensitivity or long time consumption. Therefore, it is necessary to introduce other screening indicators to assist in diagnosis and improve the efficiency of early screening of ischemic stroke.
[0003] Ischemic stroke early screening models have the advantages of high efficiency, convenience and low cost. However, the existing ischemic stroke early screening models are mainly based on clinical information such as traditional ischemic stroke risk factors and biochemical indicators, and the sensitivity and specificity of the models need to be improved.
[0004] As the executor of life activities, changes in protein abundance and function can directly reflect the physiological or pathological state of the body. Proteomics studies the composition, function and interaction of all proteins in a certain type of cell or tissue through high-throughput technology, which can provide a biological explanation for the pathological changes of ischemic stroke. Therefore, establishing an early screening model based on proteomics protein markers, identifying and monitoring protein biomarkers, will help early screening and rapid diagnosis and treatment of ischemic stroke patients.
[0005] The most widely studied method at present is to use mass spectrometry to qualitatively and quantitatively analyze proteins. This technology provides strong technical support for the detection of proteins in a large number of samples. However, in the early detection of ischemic stroke, mass spectrometry cannot meet the sensitivity requirements for the identification of specific proteins. In addition, blood is a biological sample that is easier to obtain clinically. The body secretes proteins into serum or plasma. These proteins can indicate the pathological process of the disease, but the abundance of proteins with marker significance is often extremely low, which exceeds the depth of coverage of mass spectrometry detection technology. This greatly affects the accuracy of the early screening model for ischemic stroke. Regarding the pathological mechanism of ischemic stroke, one of the more recognized views is that the immune system is involved in the occurrence and development of ischemic stroke. However, its specific mechanism is still unclear, and there is a lack of research on the changes in related protein molecules caused by the pathological process of immune response.
[0006] Therefore, a new proteomic detection technology is needed that can overcome the above-mentioned limitations and target the detection of immune response-related proteins with high sensitivity and specificity, which will not only help in the early pathological interpretation of ischemic stroke, but also realize the early screening of ischemic stroke. Summary of the invention
[0007] In view of the shortcomings of the existing technology, the present invention provides an ischemic stroke immune response-related early screening marker based on Olink proteomics, and a selection method and application thereof. The optimal protein marker combination for early screening of ischemic stroke is determined by Olink proteomics technology, effectively solving the problems of high cost, long time consumption, insufficient marker sensitivity and specificity in the existing ischemic stroke screening.
[0008] To achieve the above objectives, the technical solution of the present invention is implemented through the following technical solutions:
[0009] An early screening marker related to ischemic stroke immune response based on Olink proteomics, wherein the marker is the following 6 proteins: MASP1, STC1, HCLS1, CLEC4D, PTH1R and PIK3AP1.
[0010] The method for selecting early screening markers related to ischemic stroke immune response based on Olink proteomics includes the following steps:
[0011] S1. Clinical data collection of ischemic stroke patients and control group: Two groups of ischemic stroke patients, case group and control group, were set up as discovery set and validation set respectively;
[0012] S2. Olink proteomics detection and analysis: Blood samples were collected from the case group and control group in the above discovery set and validation set, and all Olink panels were searched and matched. The immune response panel was used for concentration detection, and then the concentrated differentially expressed proteins were identified. The receiver operating characteristic curve analysis was performed on all differentially expressed proteins, and the area under the characteristic curve was calculated.
[0013] S3. Establishment and evaluation of ischemic stroke diagnostic model: Feature variables were screened from differentially expressed proteins using LASSO and random forest, and finally variables with high credibility were determined as feature variables for model establishment and evaluated in the validation set.
[0014] Preferably, the inclusion criteria for ischemic stroke patients in S1 are: age>18 years old; first-time acute ischemic stroke patients; complete clinical data of patients, and well-preserved clinical samples; patients or guardians agree to accept the program and sign the informed consent form;
[0015] The exclusion criteria for ischemic stroke patients were as follows: patients with cardiogenic stroke; patients with cardiovascular diseases such as atrial fibrillation, myocardial infarction, heart failure, and a history of coronary artery intervention / bypass; patients with liver and kidney dysfunction; patients with mental illness; patients with malignant tumors; pregnant women, lactating women, and women of childbearing age;
[0016] The inclusion criteria for the control group were: age > 18 years old; no history of ischemic stroke; no medication in the past 3 months; agreeing to accept the program and signing the informed consent form;
[0017] The exclusion criteria for the control group were: patients with cardiovascular disease and history of coronary artery intervention / bypass grafting; patients with mental illness; patients with malignant tumors; pregnant women, lactating women and women of childbearing age.
[0018] Preferably, the Olink proteomic analysis in step S2 includes: performing Spearman correlation analysis on the expression levels of differential proteins, performing GO and KEGG enrichment analysis on the differential proteins using ggplot2; and performing interaction analysis of differential proteins using the StringDB protein interaction database.
[0019] Preferably, the method for determining the variables in step S3 is: using LASSO regression analysis, random forest analysis, and proteins with larger areas under a single characteristic curve to jointly verify the intersection and determine 6 proteins as characteristic variables to be included in the ischemic stroke diagnostic model.
[0020] Preferably, the LASSO regression analysis uses differential proteins as independent variables and utilizes LASSO regression analysis to generate a penalty function that compresses the number of variables by using the variable coefficients in the regression model.
[0021] Preferably, the top 10 characteristic variables in the random forest analysis are MASP1, STC1, HCLS1, CLEC4D, LILRB4, PTH1R, PRDX1, PIK3AP1, DCTN1, and ITGA6.
[0022] Olink proteomics-based early screening markers related to ischemic stroke immune response are used to establish early screening models for ischemic stroke, assist in clinical diagnosis of ischemic stroke, and provide targets for mechanism research.
[0023] Preferably, the early screening marker is used in the preparation of an ischemic stroke early screening kit.
[0024] The present invention provides an ischemic stroke immune response-related early screening marker based on Olink proteomics and a selection method and application thereof, which has the following advantages over the prior art:
[0025] The present invention adopts Olink proteomics based on PEA technology. When detecting each protein, a pair of specifically matched antibodies is designed to ensure the specificity of the immunological response level. By adding a one-to-one paired nucleic acid single strand to a pair of antibodies, the specificity of DNA pairing is used to avoid crosstalk at the immunological level, thereby realizing ultra-sensitive and unbiased targeted proteomics in a biological sense. It is particularly suitable for blood samples that traditional methods are not capable of. In addition, Olink proteomics technology is used to detect immune response-related proteins in the serum of ischemic stroke patients, identify the changed core proteins and their biological functions, and effectively process the high-dimensional data of omics through machine learning. The information therein is effectively mined through feature selection, dimensionality reduction and other technologies. The final determined combination of 6 protein markers can provide a biological explanation for the pathological process of the immune response, and the marker combination of 6 proteins has a high accuracy. BRIEF DESCRIPTION OF THE DRAWINGS
[0026] Figure 1 This is a profile diagram of LASSO regression dimensionality reduction coefficients of immune response-related proteins for early screening of ischemic stroke in an embodiment of the present invention;
[0027] Figure 2 This is a schematic diagram of a cross-validation curve of immune response-related proteins for early screening of ischemic stroke according to an embodiment of the present invention;
[0028] Figure 3 The top 10 proteins screened by random forest for early screening of immune response-related proteins for ischemic stroke in the embodiment of the present invention;
[0029] Figure 4 The Venn diagram of proteins screened by LASSO and random forest and proteins with AUC>0.8 in the embodiment of the present invention;
[0030] Figure 5 This is a schematic diagram of a binary logistic regression model of a training set according to an embodiment of the present invention;
[0031] Figure 6 This is a schematic diagram of a binary logistic regression model of an internal validation set of an embodiment of the present invention;
[0032] Figure 7 This is a schematic diagram of the AUC binary logistic regression model of ROC in the test set of an embodiment of the present invention;
[0033] Figure 8 A learning curve diagram of an embodiment of the present invention;
[0034] Fig. 9 This is a schematic diagram of a calibration curve according to an embodiment of the present invention;
[0035] Fig.10 This is a schematic diagram of feature importance ranking based on SHAP according to an embodiment of the present invention;
[0036] Fig.11 Attributes of the features of the embodiment of the present invention in SHAP, where red represents high feature values and blue represents low feature values;
[0037] Fig.12 Schematic diagram of SHAP forceplot for non-ischemic stroke population and ischemic stroke patients in an embodiment of the present invention, where f(x) is the model output value, the base value is the average of the model output and the training data, red represents features that push the prediction up, and blue represents features that push the prediction down;
[0038] Fig.13 is the AUC of the ROC of the validation set of the embodiment of the present invention;
[0039] Fig.14 This is a calibration curve of the model used in the verification set of the embodiment of the present invention. DETAILED DESCRIPTION
[0040] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention is clearly and completely described below in combination with the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0041] Embodiment 1:
[0042] Screening of early screening markers related to immune response in ischemic stroke based on Olink proteomics:
[0043] 1. Collection of clinical data of ischemic stroke patients and control group:
[0044] A 1:1 matching design was used to recruit patients with ischemic stroke from August 2021 to May 2023 in the Department of Neurology, Ningbo Second Hospital, China. Gender, age, and BMI were used as matching factors to match controls recruited from the physical examination center during the same period to eliminate confounding factors. The same method was used to recruit ischemic stroke cases from November 2022 to November 2023 in the Department of Neurology, Xiangcheng Third People's Hospital, Suzhou, China, and matched with controls from the physical examination center during the same period. Finally, 88 cases and 88 controls were included in the study as a validation set. The study was conducted in strict accordance with the internationally recognized principles of the Declaration of Helsinki. The protocol was approved by the Human Research Ethics Committee of Ningbo Second Hospital (approval number: YJ-NBEY-KY-2021-072-01, date: May 6, 2021). All participants received a complete explanation of the background and content of the project and provided written informed consent to participate in this study.
[0045] The inclusion criteria for patients with ischemic stroke were (1) age > 18 years old; (2) patients with first-onset acute ischemic stroke; (3) complete clinical data and well-preserved clinical samples; (4) patients or guardians agreed to accept the program and signed the informed consent. The exclusion criteria for patients with ischemic stroke were (1) patients with cardiogenic stroke; (2) patients with cardiovascular diseases such as atrial fibrillation, myocardial infarction, heart failure, and a history of coronary artery intervention / bypass; (3) patients with liver and kidney dysfunction; (4) patients with mental illness; (5) patients with malignant tumors; (6) pregnant women, breastfeeding women, and women of childbearing age. The inclusion criteria for the control group were (1) age > 18 years old; (2) no history of ischemic stroke; (3) no medication in the past 3 months; (4) agreed to accept the program and signed the informed consent. The exclusion criteria for the control group were (1) patients with cardiovascular disease and a history of coronary artery intervention / bypass grafting; (2) patients with mental illness; (3) patients with malignant tumors; and (4) pregnant women, breastfeeding women, and women of childbearing age.
[0046] Strictly trained investigators collected demographic and clinical data of the above subjects, including gender, age, height, weight, medical history, laboratory biochemical indicators, etc.
[0047] Fasting venous blood was collected from both the case group and the control group in the discovery and validation sets. The blood samples of the case group were collected within 24 hours of admission to the neurology department for ischemic stroke. Serum was obtained by centrifugation at 1800g for 10min and 13000g for 2min and immediately stored in a -80℃ refrigerator for subsequent proteomics testing.
[0048] 2. Olink proteomics detection and analysis:
[0049] use The immune response panel (Olink Proteomics AB, Uppsala, Sweden) measured the concentration of 92 proteins in the serum of cases and controls in the discovery and validation sets. The Proximity Extension Assay technology in Olink is a dual recognition immunoassay, in which two matched antibodies bind to a target simultaneously and are labeled with unique DNA oligonucleotides. Proteins in solution bring the two antibodies into proximity, allowing their DNA oligonucleotides to hybridize, serving as templates for the DNA polymerase-dependent extension step. The double-stranded DNA "barcode" so created is unique to the specific antigen and is quantitatively proportional to the initial concentration of the target protein. Hybridization and extension are followed immediately by PCR amplification, and finally the amplicons are quantified by microfluidic qPCR.
[0050] The detailed information of 92 immune response related proteins is shown in the following table:
[0051]
[0052]
[0053]
[0054]
[0055]
[0056]
[0057] The protein content was log2 processed with the standardized protein expression value to obtain the standardized protein expression value (NPX). In the discovery set, proteins with a p value less than 0.05 obtained by the T statistical test were considered differential proteins. In the discovery set, a total of 59 differentially expressed proteins were identified between the groups, of which 42 proteins were significantly upregulated in patients with ischemic stroke, while 17 proteins were significantly downregulated. Receiver operating characteristic curve (ROC) analysis was performed on all differential proteins, and the area under the ROC curve (AUC) was calculated. Spearman correlation analysis was performed on the expression levels of differential proteins, and ggplot2 was used to perform GO and KEGG enrichment analysis on differential proteins. Differential proteins were enriched in 850 GO-Terms, of which 445 were significantly different between groups. Biological processes such as adaptive immune response, innate immune response, and aging were the most significant GO-Terms. In addition, the differentially expressed proteins were enriched in 112 metabolic pathways, and 61 metabolic pathways showed significant differences between the groups. NF-kappa B signaling pathway, C-type lectin receptor signaling pathway, Intestinal immune network for IgA production and other pathways were the most significant KEGG pathways. The interaction analysis of the differentially expressed proteins was performed using the StringDB protein interaction database (http: / / string-db.org / ). IL6 and IL10 had the highest connectivity in the protein interaction network.
[0058] 3. Establishment and evaluation of ischemic stroke diagnostic model:
[0059] LASSO (Least Absolute Shrinkage and Selection Operator) and random forest were used to screen the characteristic variables from the differentially expressed proteins. At the same time, considering that the proteins with larger AUC had diagnostic significance, we compared the three methods and finally determined the variables with high credibility from each method as characteristic variables for model establishment.
[0060] The specific method is as follows: 59 differentially expressed proteins are used as independent variables, and LASSO regression analysis is used to generate a penalty function that compresses the number of variables in the regression model. When the standard error λ of the minimum distance is 0.057, 11 characteristic coefficients are non-zero among the 59 independent variables:
[0061] (ITGB6+PRDX1+ITGA6+GLB1+PTH1R+BACH1+PIK3AP1+HC LS1+CLEC4D+STC1+MASP1).
[0062] The top 10 feature variables in the random forest analysis were MASP1, STC1, HCLS1, CLEC4D, LILRB4, PTH1R, PRDX1, PIK3AP1, DCTN1, and ITGA6.
[0063] At the same time, proteins with large single AUC were considered. The intersection of these three analysis results was taken, and finally six proteins were determined as characteristic variables: MASP1, STC1, HCLS1, CLEC4D, PTH1R and PIK3AP1 were included in the ischemic stroke diagnosis model. The collinearity test showed that the variance inflation factor (VIF) of the six variables was less than 10, and there was no highly correlated relationship pair.
[0064] The 176 participants in the discovery set were split into an overall training set and a test set at an 8:2 ratio. Logistic regression was used to model the feature variables, with 5-fold cross-validation performed in the overall training set and evaluation in the test set.
[0065] The AUC, cut off value, accuracy, sensitivity, specificity, positive predictive value, negative predictive value and F1 score were calculated in the training set, internal validation set and test set respectively. If the performance of the validation set under the AUC indicator did not exceed that of the test set or the excess ratio was less than 10%, the fitting was considered successful.
[0066] The fit and stability of the model in the training set and validation set were evaluated by the learning curve, and the accuracy of the final model was evaluated by the calibration curve and verified in the external validation set.
[0067] The average AUC of the 6-protein model in the training set and internal validation set were 0.964 (95% CI: 0.929-0.999) and 0.962 (95% CI: 0.895-1.000), respectively. The AUC of the final model in the test set was 0.954 (95% CI: 0.884-1.000), and the sensitivity and specificity were 0.947 and 0.941, respectively.
[0068] Python shap 0.39.0 is used to analyze the model and show the contribution of individual features to the model and the positive or negative impact on the model.
[0069] The importance of the six features in the model is MASP1, CLEC4D, PIK3AP1, STC1, PTH1R, and HCLS1; CLEC4D, PIK3AP1, STC1, and PTH1R are positive diagnostic factors for ischemic stroke, while MASP1 is a negative diagnostic factor for ischemic stroke.
[0070] The AUC of the 6-protein combination in an independent external validation set also reached 0.857 (95% CI: 0.801-0.913), with a sensitivity and specificity of 0.864 and 0.73, respectively, indicating that the protein combination has good screening performance.
[0071] In the protein-protein interaction network, IL6 and IL10, which have the largest degree, have a certain degree of correlation in expression. CLEC4D, PIK3AP1, HCLS1 and IL6, IL10 mainly play biological functions through the C-type lectin receptorsignaling pathway (hsa04625), PI3K-Akt signaling pathway (hsa04151), and B cellreceptor signaling pathway (hsa04662), which may be the core signaling pathway that changes in the early stage of ischemic stroke.
[0072] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims
1. An early screening marker for ischemic stroke immune response based on Olink proteomics, characterized in that: The markers are the following six proteins: MASP1, STC1, HCLS1, CLEC4D, PTH1R and PIK3AP1.
2. The Olink proteomics-based early screening marker for ischemic stroke immune response according to claim 1, characterized in that: The marker selection method comprises the following steps: S1. Clinical data collection of ischemic stroke patients and control group: Two groups of ischemic stroke patients, case group and control group, were set up as discovery set and validation set respectively; S2. Olink proteomics detection and analysis: Blood samples were collected from the case group and the control group in the above discovery set and validation set, and the concentration of the protein in the serum was tested using the Olink® immune response panel. The differential proteins in the discovery set were then determined, and the receiver operating characteristic curve analysis was performed on all the differential proteins to calculate the area under the characteristic curve (AUC); S3. Establishment and evaluation of ischemic stroke diagnostic model: The intersection of LASSO, random forest and single AUC top-ranked protein was used to screen characteristic variables from the differentially expressed proteins, and finally the variables with high credibility were determined as characteristic variables for model establishment and evaluated in the validation set.
3. The Olink proteomics-based early screening marker for ischemic stroke immune response according to claim 2, characterized in that: The inclusion criteria for ischemic stroke patients in S1 are: age > 18 years old; first-onset acute ischemic stroke patients; complete clinical data of patients, and well-preserved clinical samples; patients or guardians agree to accept the program and sign the informed consent form; The exclusion criteria for ischemic stroke patients were as follows: patients with cardiogenic stroke; patients with cardiovascular diseases such as atrial fibrillation, myocardial infarction, heart failure, and a history of coronary artery intervention / bypass; patients with liver and kidney dysfunction; patients with mental illness; patients with malignant tumors; pregnant women, lactating women, and women of childbearing age; The inclusion criteria for the control group were: age > 18 years old; no history of ischemic stroke; no medication in the past 3 months; agreeing to accept the program and signing the informed consent form; The exclusion criteria for the control group were: patients with cardiovascular disease and history of coronary artery intervention / bypass grafting; patients with mental illness; patients with malignant tumors; pregnant women, lactating women and women of childbearing age.
4. The Olink proteomics-based early screening marker for ischemic stroke immune response according to claim 2, characterized in that: The Olink proteomics analysis in step S2 includes: performing Spearman correlation analysis on the expression levels of differentially expressed proteins, performing GO and KEGG enrichment analysis on the differentially expressed proteins using ggplot2; and performing interaction analysis of the differentially expressed proteins using the StringDB protein interaction database.
5. The Olink proteomics-based early screening marker for ischemic stroke immune response according to claim 2, characterized in that: The method for determining the variables in step S3 is: using LASSO regression analysis, random forest analysis, and proteins with larger areas under a single characteristic curve to jointly verify the intersection and determine 6 proteins as characteristic variables to be included in the ischemic stroke diagnosis model.
6. An application of the Olink proteomics-based early screening markers for ischemic stroke immune response as claimed in claim 1, characterized in that: The early screening marker is used in the preparation of an ischemic stroke early screening kit.
