Biomarker and application thereof

Abstract

The invention relates to a biomarker. The biomarker is selected from one or more of the following components: UBR2, NPY4R, KLRF2, HABP2, GLP1R, DR1, RUNX2, ARHGAP5, LINC02135, DDHD1, MIR4523 and TC2N. The biomarker is especially used for diagnosing cerebral apoplexy and identifying and diagnosing cerebral hemorrhage. According to the invention, 12 characteristic biomarkers are screened for model construction, the performance of the model is analyzed and evaluated by using the working characteristic (ROC) of a subject, the area under a curve (AUC) is calculated by using sklearn, the sensitivity and specificity of the model respectively reach 0.898 and 0.818, and the AUC value is 0.913; meanwhile, cerebral hemorrhage is identified and diagnosed, the sensitivity and the specificity respectively reach 0.864 and 0-854, and the AUC value is 0.909. According to the invention, the 5hmC modification spectrum of plasma evDNA is combined with a machine learning algorithm for the first time, the screened characteristic biomarker and / or model have the advantages of strong specificity and high sensitivity, and the problems of low efficiency and poor accuracy in the cerebral apoplexy and cerebral hemorrhage identification process in the prior art are solved.

Description

Technical Field

[0001] The present invention relates to the technical field of molecular biology, and in particular to a biomarker and applications thereof. Background Art

[0002] Stroke is the second leading non-communicable cause of death worldwide, and its classification and treatment are highly dependent on early and accurate identification. Patients with large vessel occlusion (LVO-AIS) in acute ischemic stroke (AIS) require urgent intravascular thrombectomy, while patients with hemorrhagic stroke (such as intracranial hemorrhage, ICH) need to avoid anticoagulation and control blood pressure. However, current clinical triage technology has significant bottlenecks. Although CT / MRI is the "gold standard" for diagnosis, the low accessibility of equipment in the pre-hospital setting and the time-consuming transportation have led to a severe compression of the treatment time window (such as the 4.5-hour intravenous thrombolysis window for AIS).

[0003] CN119470922A discloses biomarkers for detecting traumatic brain injury, including NSE, GFAP, and S100B. However, these biomarkers can only be used to predict traumatic brain injury and cannot distinguish between acute ischemic and hemorrhagic strokes. CN110257525A discloses biomarkers with significant potential for tumor diagnosis and their uses. This method utilizes methylation-modified tumor markers as a tumor diagnosis approach, but this method relies on expression level identification, making it difficult to meet the demand for rapid identification.

[0004] Recent studies have demonstrated that extracellular vesicles (EVs) offer unique advantages in early disease diagnosis due to their ability to carry parent cell-specific biomolecules (e.g., DNA and RNA). In particular, DNA within EVs (evDNA) not only retains the epigenetic signatures of the parental cells (e.g., 5-hydroxymethylcytosine, 5hmC) but also exhibits significantly greater stability than free DNA (EV-DNA). 5hmC, a product of 5-methylcytosine (5mC) oxidation by TET enzymes, is closely associated with gene transcriptional activation, and its dynamic changes have been shown to be highly correlated with the pathological progression of ischemic brain injury. However, the use of evDNA's 5hmC epigenetic signature for noninvasive and rapid identification of stroke types remains a mystery.

[0005] In addition, on the one hand, there are differences in understanding among those skilled in the art; on the other hand, the applicant studied a large number of documents and patents when making the present invention, but due to space limitations, not all details and contents are listed in detail. However, this does not mean that the present invention does not have the characteristics of these prior arts. On the contrary, the present invention already has all the characteristics of the prior art, and the applicant reserves the right to add relevant prior art to the background technology. Summary of the Invention

[0006] In response to the deficiencies of the prior art, the first aspect of the present invention provides a biomarker selected from one or more of the following: UBR2, NPY4R, KLRF2, HABP2, GLP1R, DR1, RUNX2, ARHGAP5, LINC02135, DDHD1, MIR4523, and TC2N.

[0007] According to a preferred embodiment, the biomarker is the presence and / or content of a biomarker detected in a body fluid sample of an individual.

[0008] According to a preferred embodiment, the body fluid sample can be any one of blood, plasma, urine, saliva, and sweat. Preferably, the body fluid sample of the individual being tested is peripheral blood. Preferably, the body fluid sample of the individual being tested is plasma.

[0009] The second aspect of the present invention provides the use of the biomarker provided in the first aspect of the present invention in preparing products and / or models for diagnosing stroke and differentially diagnosing cerebral hemorrhage.

[0010] A third aspect of the present invention provides a kit for diagnosing stroke and differentially diagnosing cerebral hemorrhage, comprising a detection reagent for detecting the biomarker provided in the first aspect of the present invention. The biomarker is selected from one or more of the following: UBR2, NPY4R, KLRF2, HABP2, GLP1R, DR1, RUNX2, ARHGAP5, LINC02135, DDHD1, MIR4523, and TC2N.

[0011] The fourth aspect of the present invention provides a model for diagnosing stroke and differentially diagnosing cerebral hemorrhage, wherein the input variable of the model is the content of the biomarker provided by the first aspect of the present invention.

[0012] According to a preferred embodiment, the method for determining the content of the biomarker is 5hmC high-throughput detection. Specifically, the method for determining the content of the biomarker is 5hmC-Seal.

[0013] According to a preferred embodiment, a model for diagnosing stroke and differentially diagnosing cerebral hemorrhage is obtained by the following construction method, which comprises the following steps:

[0014] Plasma samples from multiple stroke and cerebral hemorrhage patients were tested to obtain 5hmC sequencing data of extracellular vesicle DNA in plasma, and the two sets of data were intersected;

[0015] The obtained sequencing data is sequentially subjected to first filtering, screening, and second filtering to obtain biomarkers for diagnosing stroke and differentially diagnosing cerebral hemorrhage;

[0016] The first filtering includes: removing peak information that only appears in less than or equal to 10 samples,

[0017] The screening process included: using DEseq2 software to align sequencing data from different samples, retaining 5hmC peak regions with more than 50 reads, and identifying differential biomarkers of 5hmC upregulation and downregulation according to |log2FoldChange|>=0.5 and p value<0.01.

[0018] The second filtering includes: filtering the differential biomarkers obtained by the screening using the recursive feature elimination algorithm (RFECV) in Scikit-Learn to obtain biomarkers for diagnosing stroke and differentially diagnosing cerebral hemorrhage: UBR2, MIR4523, NPY4R, KLRF2, GLP1R, RUNX2, LINC02135, ARHGAP5, TC2N, DR1, HABP2 and DDHD1;

[0019] The obtained biomarker data is input into a machine learning model, the model is trained, and the trained model is stored to obtain a model for diagnosing stroke and differentially diagnosing cerebral hemorrhage.

[0020] According to a preferred embodiment, in the second filtering step, the parameters used are: (estimator=LogisticRegressionCV(class_weight='balanced', cv=2, max_iter=1000), scoring='accuracy').

[0021] According to a preferred embodiment, the machine learning model includes: training a logistic regression CV model (LR).

[0022] A fifth aspect of the present invention provides a method for diagnosing stroke and differentially diagnosing cerebral hemorrhage, which is implemented based on the model provided in the fourth aspect of the present invention and comprises the following steps:

[0023] Collecting samples to be tested from subjects;

[0024] Detecting 5hmC sequencing data of extracellular vesicle DNA in the test sample to obtain the content of the biomarker provided by the first aspect of the present invention;

[0025] The obtained biomarker levels are input into the model provided by the fourth aspect of the present invention.

[0026] According to a preferred embodiment, the sample to be tested is peripheral blood or plasma.

[0027] The present invention also provides the use of a reagent for detecting the following proteins in the preparation of a product for diagnosing stroke and differentially diagnosing cerebral hemorrhage, wherein the proteins are selected from one or more of the following: UBR2, NPY4R, KLRF2, HABP2, GLP1R, DR1, RUNX2, ARHGAP5, LINC02135, DDHD1, MIR4523, and TC2N.

[0028] Technical effect of the present invention: The present invention screened 12 characteristic biomarkers for model construction, adopted receiver operating characteristic (ROC) analysis to evaluate the performance of the model, and used sklearn to calculate the area under the curve (AUC). The sensitivity and specificity of the model reached 0.898 and 0.818, respectively, and the AUC value was 0.913. At the same time, for differential diagnosis of cerebral hemorrhage, the sensitivity and specificity reached 0.864 and 0.854, respectively, and the AUC value was 0.909. The present invention combines the 5hmC modification spectrum of plasma evDNA with a machine learning algorithm for the first time. The screened characteristic biomarkers and / or models have the advantages of strong specificity and high sensitivity, and can provide an accurate non-invasive diagnostic method at the molecular level for stroke and cerebral hemorrhage. The diagnostic model is reliable and the diagnostic results are highly accurate. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 The experimental design process provided by the present invention randomly divides clinical samples into a training group, a validation group, and an identification group;

[0030] Figure 2 Related parameters of 12 5hmC characteristic biomarkers in the model for diagnosing stroke and differentially diagnosing cerebral hemorrhage screened by the present invention;

[0031] Figure 3 Receiver operating characteristic (ROC) curves of the diagnostic model constructed using 12 5hmC signature biomarkers in the training, validation, and identification groups (Figures AC).

[0032] Figure 4 The diagnostic accuracy results of individual markers among the 12 5hmC signature biomarkers in the training set;

[0033] Figure 5 To validate the diagnostic accuracy of individual markers among the 12 5hmC signature biomarkers in the panel;

[0034] Figure 6 To identify the diagnostic accuracy of individual markers among the 12 5hmC signature biomarkers in the panel;

[0035] Figure 7 The structural formula is DBCO-PEG4-Biotin. DETAILED DESCRIPTION

[0036] The present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be understood that these descriptions are exemplary only and are not intended to limit the scope of the present disclosure. In the detailed description below, for ease of explanation, many specific details are set forth to provide a comprehensive understanding of the embodiments of the present disclosure. However, it is obvious that one or more embodiments can also be implemented without these specific details. In addition, in the following description, descriptions of well-known technologies are omitted to avoid unnecessary confusion of the concepts of the present disclosure. In order to make the purposes, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be described clearly and completely below. Where specific conditions are not specified in the embodiments, they are carried out according to conventional conditions or conditions recommended by the manufacturer. All terms (including technical and scientific terms) used herein have the meanings commonly understood by those skilled in the art, unless otherwise defined. It should be noted that the terms used here should be interpreted as having meanings consistent with the context of this specification and should not be interpreted in an idealized or overly rigid manner.

[0037] Stroke includes ischemic stroke and hemorrhagic stroke. Cerebral hemorrhage specifically refers to a type of hemorrhagic stroke, in which blood vessels in the brain parenchyma rupture and bleed.

[0038] The present invention provides the following technical methods.

[0039] Step 1: Extracellular vesicle isolation from plasma samples

[0040] 1. Collect whole blood (8 mL) and place in a 10 mL Kangwei (CW2815M) cfDTM cell-free nucleic acid collection tube (operate at room temperature and store at room temperature (18-25°C) after blood collection. Plasma extraction must be performed within 72 hours at most).

[0041] 2. Centrifuge for the first time at 1350g at 4°C for 12 min. Carefully remove the pale yellow supernatant (removing blood cells and platelets) and transfer it to a 2 mL DNase-free sterile centrifuge tube.

[0042] 3. Centrifuge for the second time at 1350g at 4°C for 12 min. Carefully remove the supernatant (to remove white blood cells and cell debris) and transfer to two to three 2 mL DNase-free sterile centrifuge tubes.

[0043] 4. Centrifuge for the third time at 13500g at 4°C for 5 min. Carefully remove the supernatant and transfer it to a 2 mL DNase-free sterile centrifuge tube and freeze it at -80°C.

[0044] Step 2: Isolation of plasma extracellular vesicles (EVs)

[0045] EVs were isolated from plasma samples using the Wayen Exosome Isolation Kit (EIQ3-02001).

[0046] Step 3: Plasma EV-DNA extraction

[0047] 1. Use the Quick-DNA Miniprep Plus Kit (ZYMO, D4069) to extract extracellular vesicle DNA (EV-DNA) from the extracellular vesicle (EV) sample prepared in step 1.

[0048] 2. EV-DNA concentration was determined using Qubit 3.0.

[0049] Step 4: Fill the ends of EV-DNA and connect them to sequencing adapters

[0050] According to the instructions of KAPA HyperPlus Library Preparation Kit (KK8514), perform the following steps:

[0051] 1. Prepare a reaction mixture containing 20 μL of 10 ng EV-DNA (obtained in step 2 above), 2-8 μL End Repair & A-Tailing Buffer, and 1.2 μL End Repair & A-Tailing Enzyme mix (total volume 24 μL), as shown in Table 1.

[0052] Table 1

[0053] reactants Volume (μL) EV-DNA 20 End Repair & A-Tailing Buffer 2.8 End Repair&A-Tailing Enzyme Mix 1.2

[0054] The reaction mixture prepared according to Table 1 was incubated at 20°C for 30 minutes and then at 65°C for 30 minutes, as shown in Table 2.

[0055] Table 2

[0056]

[0057] 2. Next, prepare the following ligation reaction mixture in a 1.5 mL low-binding EP tube: 2 μL Nuclease Free Water, 12 μL Ligation Buffer, and 4 μL DNA Ligase, as shown in Table 3.

[0058] Table 3

[0059] reactants Volume (μL) Nuclease-free water 2 Ligation Buffer 12 DNA Ligase 4

[0060] 2 μL of sequencing index (PKR2015, PKR2016, and PKR2017) was added to 18 μL of ligation reaction mixture, mixed, added to 24 μL of reaction sample, and heated at 20°C for 4 hours, as shown in Table 4.

[0061] Table 4

[0062] step Temperature (℃) time Adaptor Ligation 20 4 hours HOLD 4 ∞

[0063] 3. Use DNA Clean & Concentrator 5 (ZYMO, D4014) purification kit to purify the reaction product and elute with 20 μL elution buffer to obtain the final DNA ligation sample.

[0064] Step 5: 5hmC labeling

[0065] Prepare a 4 μL labeling reaction mixture: 1 μL 50 μM UDP-6-N3-Glu (Liaoning Cook Biotechnology Co., Ltd., Cat. No. 537039-67-1), 2.5 μL βGTase (NEB), and 2.5 μL HEPES buffer (pH 8.0, final concentration 50 mM), as shown in Table 5. Add 6 μL of the labeling reaction mixture to 20 μL of the DNA ligation sample. Incubate the mixture in a 37°C water bath for 1 hour.

[0066] Table 5

[0067] reactants Volume (μL) hmC-labeled substrate 1 1 β-GTase 2.5 HEPES buffer 2.5

[0068] The mixed solution was removed and the reaction product was purified using a DNA Clean & Concentrator 5 (ZYMO, D4014) purification kit to obtain 30 μL of purified DNA;

[0069] Then, 1 μL of 45 μM DBCO-PEG4-Biotin (Click Chemistry Tools) was added to 30 μL of purified DNA and incubated in a 37°C water bath for 1 h;

[0070] The structural formula of DBCO-PEG4-Biotin is as follows Figure 7 shown.

[0071] The reaction product was then purified using a DNA Clean & Concentrator 5 (ZYMO, D4014) purification kit to obtain 30 μL of purified labeled product.

[0072] Step 6: 5hmC enrichment

[0073] Equilibrate the bound magnetic beads as follows: remove 2.5 μL of Dynabeads (Invitrogen, 65306) and add 100 μL of wash buffer (5 mM Tris (pH 7.5), 1 M NaCl, and 0.02% Tween 20), mix by pipetting, place on a 1.5 mL magnetic stand, wash the bound magnetic beads three times with 100 μL of wash buffer, and finally add 100 μL of wash buffer, mix the magnetic beads, and vortex for 30 min;

[0074] After 30 min, the magnetic beads were washed three times with 100 μL of washing buffer, and finally 32 μL of binding buffer (10 mM Tris (pH 7.5), 2 M NaCl, and 0.04% Tween20) was added and mixed well;

[0075] Next, the purified labeled product obtained in step 4 above was added to the magnetic bead mixture and mixed in a rotating mixer for 30 minutes to allow for full binding;

[0076] Finally, the magnetic beads were washed five times with 100 μL of wash buffer and 23.8 μL of RNase-Free water was added.

[0077] Step 7: PCR amplification

[0078] 1. Add 25 μL of 2×PCR mastermix and 1.25 μL of PCR primers to the purified labeled product obtained in the final system of step 5 above (total volume of 50 μL), and amplify according to the temperature and conditions of the PCR reaction cycle shown in Table 6.

[0079] Table 6

[0080]

[0081] 2. The amplified product was purified using AmpureXP beads (KK8001) and finally eluted with 20 μL of elution buffer to obtain the final 5hmC library. The library concentration was then determined using Qubit 3.0.

[0082] Step 8: High-throughput sequencing after quality control of the 5hmC library

[0083] 1. The obtained 5hmC library was quality controlled using the Fragment Analyzer™ fully automated capillary electrophoresis system (reagent kit: DNF-900; software used: Fragment Analyzer instrument control software, PROSize data analysis software) to determine the size of the DNA fragments in the library and whether it contains impurities (the library size is approximately 300 bp).

[0084] 2. Then, qPCR concentration determination was performed (kit: KAPA SYBR FAST Universal qPCR Kit (KK4601)) to determine whether the sample library met the sequencing criteria. The specific steps provided by the present invention are as follows.

[0085] Preparation sample:

[0086] Prepare five 1.5 mL EP tubes, mixing four samples to be sequenced into one tube, for a total of 20 samples (indexes cannot be repeated). Pipet 5 ng of each 5hmC library sample, with a total volume of 20 μL per tube, for a final concentration of 1 ng / μL.

[0087] Sample reaction system (20 μL): 4 μL of the library in each EP tube was aspirated for qPCR quantification, as shown in Table 7.

[0088] Table 7

[0089]

[0090]

[0091] The qPCR program settings are shown in Table 8.

[0092] Table 8

[0093]

[0094] Result analysis: Analyze using qPCR operating software to determine whether the 5hmC library is degraded and whether it meets the sequencing requirements.

[0095] 3. The library (16 μL) that passed the quality inspection was sequenced using Illumina NovaSeq6000. The sequencing throughput of each sample was 4-6G, and the sequencing band size was 150 bp at both ends.

[0096] Step 9: Alignment of raw sequencing data

[0097] 1. Each raw sequencing FASTQ data was first trimmed for low-quality data using Trimmomatic software, and then aligned to the human genome hg19 using Bowtie2 software;

[0098] 2. Use MACS software to identify reads containing 5hmC peaks with the following parameters: effective genome size = 2.72e+09; tag size = 38; band width = 100; model fold = 10; P value cutoff = 1.00e-05, and use this to call peaks and generate counts files;

[0099] 3. Use DEseq2 software to compare counts files from different samples, and screen out differential biomarkers with up- and down-regulation of 5hmC according to |log2FoldChange|>=0.5 and pvalue<0.05.

[0100] The present invention provides a biomarker, which is particularly useful for diagnosing stroke and identifying cerebral hemorrhage. This embodiment includes 279 samples, including 131 patients with stroke, 48 patients with cerebral hemorrhage, and 100 healthy individuals. Peripheral blood samples (3-4 mL) were taken from the above 279 patients, and extracellular vesicle DNA was extracted from the plasma for 5hmC-Seal high-throughput sequencing (5hmC library construction and high-throughput sequencing refer to the "Technical Methods" in the specific examples section). The records of all participants have complete medical history, clinical and biochemical parameter information. At the same time, the samples were randomly divided into a training cohort, a validation cohort, and an identification group according to a certain ratio, such as Figure 1 shown.

[0101] Example 1

[0102] In the training group, the DEseq2 package was used to perform differential analysis between stroke patients and healthy controls, and 5hmC-specific biomarkers were screened according to |log2FoldChange|>=0.5 and p value<0.05. Further, a machine learning method was used to screen characteristic markers (using parameters: estimator=LogisticRegressionCV(class_weight='balanced', cv=2, max_iter=1000), scoring='accuracy'), and 12 biomarkers were obtained, including: UBR2, NPY4R, KLRF2, HABP2, GLP1R, DR1, RUNX2, ARHGAP5, LINC02135, DDHD1, MIR4523 and TC2N, as shown in Figure 2. Figure 2 shown.

[0103] Figure 3Receiver operating characteristic (ROC) curves of the diagnostic model constructed using 12 5hmC signature biomarkers in the training, validation, and identification groups. In this example, a stroke diagnosis model was constructed using the 12 screened characteristic markers, and the performance of the model was evaluated by receiver operating characteristic (ROC) analysis. The AUC value of the model diagnosis was 0.975, as shown in Figure A. At the same time, in the training group, this example analyzed the diagnostic performance of each of the 12 markers, including ARHGAP5 (AUC = 0.58), DDHD1 (AUC = 0.60), DR1 (AUC = 0.73), GLP1R (AUC = 0.69), HABP2 (AUC = 0.65), LINC02135 (AUC = 0.59), MIR4523 (AUC = 0.58), NPY4R (AUC = 0.76), RUNX2 (AUC = 0.50), TC2N (AUC = 0.56), UBR2 (AUC = 0.79), and KLRF2 (AUC = 0.79). Figure 4 Among the results, UBR2 (AUC = 0.79) and KLRF2 (AUC = 0.79) showed the highest diagnostic accuracy.

[0104] Example 2

[0105] In order to test the diagnostic ability of the stroke diagnosis model, this example uses a validation group to test the diagnostic efficacy of the model. The constructed diagnostic model is used to diagnose stroke patients in the validation group. This example found that the AUC value of the model in the validation group was as high as 0.913. Figure 3 As shown in Figure B; At the same time, in the validation group, the diagnostic accuracy of individual markers among the 12 characteristic 5hmC biomarkers, among which NPY4R (AUC = 0.83) and KLRF2 (AUC = 0.73) showed the highest diagnostic accuracy, as shown in Figure 2. Figure 5 The diagnostic accuracy of the 12 signature 5hmC biomarkers in the validation set was as follows: ARHGAP5 (AUC = 0.53); DDHD1 (AUC = 0.57); DR1 (AUC = 0.67); GLP1R (AUC = 0.54); HABP2 (AUC = 0.53); LINC02135 (AUC = 0.59); MIR4523 (AUC = 0.54); RUNX2 (AUC = 0.50); ​​TC2N (AUC = 0.65); and UBR2 (AUC = 0.68).

[0106] Example 3

[0107] In order to test the ability of the model to differentiate diagnosis, this example tests the ability of the stroke diagnosis model in the identification group. The constructed diagnostic model is used to diagnose stroke patients in the identification group. This example found that the AUC value of the model in the identification group is as high as 0.909. Figure 3 As shown in Figure C.

[0108] The diagnostic accuracy of individual markers among the 12 characteristic 5hmC biomarkers in the identification group was as follows: ARHGAP5 (AUC = 0.66), LINC02135 (AUC = 0.66), and MIR4523 (AUC = 0.66). The diagnostic accuracy of the other 9 markers was as follows: DDHD1 (AUC = 0.63); DR1 (AUC = 0.65); GLP1R (AUC = 0.63); HABP2 (AUC = 0.60); KLRF2 (AUC = 0.64); NPY4R (AUC = 0.63); RUNX2 (AUC = 0.63); TC2N (AUC = 0.63); UBR2 (AUC = 0.61), Figure 6 shown.

[0109] Therefore, the diagnostic model constructed based on the above 12 5hmC characteristic biomarkers provided in this embodiment can effectively diagnose stroke patients.

[0110] It should be noted that the above-mentioned specific embodiments are exemplary, and those skilled in the art can come up with various solutions inspired by the disclosure of the present invention, and these solutions also belong to the disclosure scope of the present invention and fall within the protection scope of the present invention. Those skilled in the art should understand that the present invention specification and its drawings are illustrative and do not constitute a limitation on the claims. The scope of protection of the present invention is defined by the claims and their equivalents. The present invention specification contains multiple inventive concepts, such as "according to a preferred embodiment", which means that the corresponding paragraph discloses an independent concept, and the applicant reserves the right to file a divisional application based on each inventive concept. Throughout the text, the features guided by "according to a preferred embodiment" are only an optional method and should not be understood as a must-have setting. Therefore, the applicant reserves the right to abandon or delete the relevant preferred features at any time.

Claims

1. A biomarker for diagnosing stroke and differentially diagnosing cerebral hemorrhage, characterized in that: The biomarker is selected from one or more of the following: UBR2, NPY4R, KLRF2, HABP2, GLP1R, DR1, RUNX2, ARHGAP5, LINC02135, DDHD1, MIR4523, and TC2N.

2. The biomarker according to claim 1, characterized in that The biomarkers are the presence and / or content of biomarkers detected in a body fluid sample of an individual.

3. The biomarker according to claim 2, characterized in that The body fluid sample can be any one of blood, plasma, urine, saliva, and sweat. Preferably, the body fluid sample of the test individual is peripheral blood or plasma.

4. Use of the biomarker according to any one of claims 1 to 3 in the preparation of products and / or models for diagnosing stroke and differentially diagnosing cerebral hemorrhage.

5. Use of a reagent for detecting the following proteins in the preparation of a product for diagnosing stroke and differentially diagnosing cerebral hemorrhage, characterized in that: The protein is selected from one or more of the following: UBR2, NPY4R, KLRF2, HABP2, GLP1R, DR1, RUNX2, ARHGAP5, LINC02135, DDHD1, MIR4523, and TC2N.

6. A kit for diagnosing stroke and differentially diagnosing cerebral hemorrhage, characterized in that: The kit includes a detection reagent for detecting a biomarker, wherein the biomarker is selected from one or more of the following: UBR2, NPY4R, KLRF2, HABP2, GLP1R, DR1, RUNX2, ARHGAP5, LINC02135, DDHD1, MIR4523, and TC2N.

7. A model for diagnosing stroke and differentially diagnosing cerebral hemorrhage, characterized in that: The input variable of the model is the content of the biomarker, wherein the biomarker is selected from one or more of the following: UBR2, NPY4R, KLRF2, HABP2, GLP1R, DR1, RUNX2, ARHGAP5, LINC02135, DDHD1, MIR4523, and TC2N.

8. The model according to claim 7, characterized in that The model is obtained by the following construction method, which includes the following steps: Plasma samples from multiple stroke and cerebral hemorrhage patients were tested to obtain 5hmC sequencing data of extracellular vesicle DNA in plasma, and the two sets of data were intersected; The obtained sequencing data is sequentially subjected to first filtering, screening, and second filtering to obtain biomarkers for diagnosing stroke and differentially diagnosing cerebral hemorrhage; The obtained biomarker data is input into a machine learning model, the model is trained, and the trained model is stored to obtain a model for diagnosing stroke and differentially diagnosing cerebral hemorrhage.

9. The model according to claim 8, characterized in that The first filtering included removing peak information that appeared in less than or equal to 10 samples. The screening included aligning sequencing data from different samples using DEseq2 software, retaining 5hmC peak regions with more than 50 reads, and identifying differential biomarkers of 5hmC upregulation and downregulation according to |log2FoldChange|>=0.5 and p value<0.

01. The second filtering included filtering the differential biomarkers obtained through the screening using the recursive feature elimination algorithm (RFECV) in Scikit-Learn to obtain biomarkers for diagnosing stroke and differentially diagnosing cerebral hemorrhage: UBR2, MIR4523, NPY4R, KLRF2, GLP1R, RUNX2, LINC02135, ARHGAP5, TC2N, DR1, HABP2, and DDHD1.

10. The model according to claim 8 or 9, characterized in that In the second filtering step, the parameters used are: (estimator=LogisticRegressionCV(class_weight='balanced', Cv=2, max_iter=1000), scoring='accuracy').

Images