Use of protein markers in the manufacture of a product for diagnosing parkinson or predicting parkinson's disease

Abstract

The application provides application of a protein marker in preparation of a product for diagnosing Parkinson or predicting Parkinson disease. The protein marker comprises CPA2 and / or DPEP1. The application solves the problem of poor effect of a product for diagnosing or predicting Parkinson disease in the prior art and is suitable for the Parkinson disease diagnosis product application field.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of Parkinson's disease diagnosis products, in particular, to application of a protein marker in preparation of a product for diagnosing Parkinson's disease or predicting Parkinson's disease. BACKGROUND

[0002] Parkinson's disease (PD) is a common neurodegenerative disease in the elderly, and its main clinical features are tremor, muscle rigidity and reduced movement. As one of the most common neurological diseases, the earliest symptoms of PD are often difficult to detect and are easily overlooked. For a long time, the main diagnostic method of PD has relied on personal clinical diagnosis (routine examination, changes in abnormal electrocardiogram of the disease itself, etc.), motor function assessment (HY or MDS-UPDRS score is calculated by investigating the list), and neurological examination. There is no blood or laboratory test to diagnose non-hereditary Parkinson's disease, which means that patients need to have certain physical symptoms to determine the disease. Due to the lack of decisive screening or experiments, early Parkinson's patients may not meet the clinical diagnostic criteria, and this lack of specificity also means that patients may be misdiagnosed with Parkinson's disease, resulting in missing the best period for treatment.

[0003] The lack of effective early diagnosis methods is the main factor for poor prognosis in the middle and late stages of PD patients. Although the U.K.’s Parkinson’s Disease Society Brain Bank is the gold standard for diagnosing PD, and the Parkinson's Disease Diagnostic Rating Scale (MDS-UPDRS) has gradually become a more mainstream clinical diagnostic standard, the rating scale mainly relies on the subjective cognition, mental state and motor symptoms of the patient, which limits its application in early screening.

[0004] Imaging diagnosis is difficult to confirm, and DaTscan imaging scanning approved by the U.S. Food and Drug Administration in 2011 can allow doctors to view images of the brain's dopamine system. DaTscan involves injecting a radioactive drug and binding to dopamine neurotransmitters in the brain (one of the reasons for Parkinson's is the lack of dopamine). However, the results of the scan cannot show whether or not to have Parkinson's, and the cost is high and makes patients receive a high dose of radiation every year.

[0005] In recent years, biomolecular diagnosis is widely used for early screening and auxiliary diagnosis of diseases, and improves the survival rate of patients. Blood-cerebrospinal fluid detection can detect the level of dopamine; the content of metabolites such as 5-hydroxytryptamine and indole acetic acid is reduced; dopamine beta hydroxylase is reduced; somatostatin is significantly reduced; and the level of gamma-aminobutyric acid is reduced. However, the content of patient metabolites is greatly affected by individual factors such as environmental factors, living habits and eating habits, and there are problems such as insufficient sensitivity or specificity of the marker, lack of a large number of population samples for verification, and limitation on clinical application. We need more research to understand biological complexity, and it is crucial to find more effective molecular diagnostic markers.

[0006] Other neurological diseases can cause symptoms similar to Parkinson's disease. The term Parkinsonism is used to refer to a patient with symptoms similar to Parkinson's disease. About 15% of Parkinson's disease is caused by other diseases. In addition to genetic causes found in a small number of cases, Parkinson-like diseases have no cure or known causes, such as progressive supranuclear palsy (PSP), multiple system atrophy (MSA), and Lewy body dementia. Therefore, it is urgent to find biomarkers that can be used to diagnose early-stage patients with PD and distinguish Parkinson's disease in order to perform personalized treatment for patients. SUMMARY

[0007] The main purpose of the present application is to provide an application of a protein marker in the preparation of a product for diagnosing Parkinson's disease or predicting Parkinson's disease, so as to solve the problem of poor effect of the product for diagnosing or predicting Parkinson's disease in the prior art.

[0008] In order to achieve the above-mentioned purpose, according to a first aspect of the present application, an application of a protein marker in the preparation of a product for diagnosing Parkinson's disease or predicting Parkinson's disease is provided, wherein the protein marker comprises CPA2 and / or DPEP1.

[0009] Further, the protein marker further comprises any one or more of the following: CXCL6, IL17C, CCL19, CCL27, GPNMB, FGF23, IL32, PSME1 or TBCB; preferably, the protein marker further comprises any one or more of CX3CL1, CLEC10A, IL10, NTRK3, SNCG, PMVK, PTPN.

[0010] Further, the protein markers are selected from any one of the following groups: (1) CCL19, CX3CL1, IL10, CPA2, and DPEP1; or (2) CCL19, CX3CL1, IL.17C, CPA2, and DPEP1; or (3) CCL19, IL10, CPA2, NTRK3, and DPEP1; or (4) CCL19, CX3CL1, CPA2, and DPEP1; or (5) CX3CL1, IL10, CPA2, and DPEP1; or (6) CX3CL1, CXCL6, IL.17C, CPA2, and PMVK; or (7) CX3CL1, CXCL6, IL.17C, CPA2, and PTPN1; or (8) CX3CL1, CXCL6, CPA2, GPNMB, and PTPN1; or (9) CX3CL1, CXCL6, CPA2, GPNMB, and PMVK; or (10) CX3CL1, CXCL6, IL.17C, CPA2, and TBCB; or (11) CLEC10A, CPA2, NTRK3, and SNCG; or (12) CLEC10A, CPA2, NTRK3, and DPEP1; or (13) CLEC10A, CPA2, NTRK3, DPEP1, and SNCG; or (14) IL10, CLEC10A, CPA2, NTRK3, and SNCG; or (15) CLEC10A, NTRK3, and DPEP1; or (16) CX3CL1, CXCL6, IL10, CLEC10A, and DPEP1; or (17) CX3CL1, CXCL6, IL.17C, CLEC10A, and DPEP1; or (18) CX3CL1, CXCL6, CLEC10A, and DPEP1; or (19) CX3CL1, CXCL6, CLEC10A, DPEP1, and PSME1; or (20) CX3CL1, CXCL6, FGF.23, CLEC10A, and DPEP1; or (21) IL10, CCL27, and DPEP1; or (22) IL10, CPA2, and DPEP1; or (23) IL10, CCL27, DPEP1, and IL32; or (24) IL10, CLEC10A, CPA2, and DPEP1; or (25) IL10, CLEC10A, and DPEP1; or (26) CPA2, and DPEP1; or (27) CPA2, DPEP1, and CX3CL1; or (28) CPA2, DPEP1, CX3CL1, and CLEC10A; or (29) CPA2, DPEP1, CX3CL1, CLEC10A, and IL10.

[0011] Further, the application comprises detecting the content of the protein marker by using a detection reagent; preferably, the method for detecting the content of the protein marker comprises Olink technology, immunoblotting experiment or enzyme-linked immunosorbent assay; preferably, the detection reagent is an antibody of the protein marker; preferably, the antibody is disposed on a solid support; preferably, the solid support comprises an enzyme-labeled plate, a membrane support or a microsphere; preferably, the membrane support is selected from a nitrocellulose membrane, a glass cellulose membrane or a nylon membrane; preferably, the antibody is a monoclonal antibody or a polyclonal antibody; preferably, the related product comprises a kit or an electronic prediction device.

[0012] In order to achieve the above-mentioned object, according to a second aspect of the present application, a kit for diagnosing or predicting Parkinson's disease in a subject is provided, which comprises a detection reagent for detecting the content of a protein marker, wherein the protein marker comprises CPA2 and / or DPEP1.

[0013] Further, the protein markers further include any one or more of: CXCL6, IL17C, CCL19, CCL27, GPNMB, FGF23, IL32, PSME1, or TBCB; preferably, the protein markers further include any one or more of CX3CL1, CLEC10A, IL10, NTRK3, SNCG, PMVK, PTPN; preferably, the protein markers include: (1) CCL19, CX3CL1, IL10, CPA2, and DPEP1; or (2) CCL19, CX3CL1, IL.17C, CPA2, and DPEP1; or (3) CCL19, IL10, CPA2, NTRK3, and DPEP1; or (4) CCL19, CX3CL1, CPA2, and DPEP1; or (5) CX3CL1, IL10, CPA2, and DPEP1; or (6) CX3CL1, CXCL6, IL.17C, CPA2, and PMVK; or (7) CX3CL1, CXCL6, IL.17C, CPA2, and PTPN1; or (8) CX3CL1, CXCL6, CPA2, GPNMB, and PTPN1; or (9) CX3CL1, CXCL6, CPA2, GPNMB, and PMVK; or (10) CX3CL1, CXCL6, IL.17C, CPA2, and TBCB; or (11) CLEC10A, CPA2, NTRK3, and SNCG; or (12) CLEC10A, CPA2, NTRK3, and DPEP1; or (13) CLEC10A, CPA2, NTRK3, DPEP1, and SNCG; or (14) IL10, CLEC10A, CPA2, NTRK3, and SNCG; or (15) CLEC10A, NTRK3, and DPEP1; or (16) CX3CL1, CXCL6, IL10, CLEC10A, and DPEP1; or (17) CX3CL1, CXCL6, IL.17C, CLEC10A, and DPEP1; or (18) CX3CL1, CXCL6, CLEC10A, and DPEP1; or (19) CX3CL1, CXCL6, CLEC10A, DPEP1, and PSME1; or (20) CX3CL1, CXCL6, FGF.23. CLEC10A and DPEP1; or (21) IL10, CCL27 and DPEP1; or (22) IL10, CPA2 and DPEP1; or (23) IL10, CCL27, DPEP1 and IL32; or (24) IL10, CLEC10A, CPA2 and DPEP1; or (25) IL10, CLEC10A and DPEP1; or (26) CPA2 and DPEP1; or (27) CPA2, DPEP1 and CX3CL1; or (28) CPA2, DPEP1, CX3CL1 and CLEC10A; or (29) CPA2, DPEP1, CX3CL1, CLEC10A and IL10.

[0014] To achieve the above object, according to a third aspect of the present application, there is provided an electronic device for diagnosing or predicting Parkinson's disease, comprising: an acquisition detection module configured to acquire a biological sample, detect the biological sample, and obtain the content of protein markers in the biological sample; a diagnosis or prediction module configured to establish a regression equation using a logistic regression model, a support vector machine model or a naive Bayes model and the content of the protein markers, construct a model, and then output a prediction result according to the model; preferably, the biological sample is a plasma sample; further, the protein markers in the biological sample include CPA2 and / or DPEP1; preferably, the protein markers in the biological sample further include any one or more of the following: CXCL6, IL17C, CCL19, CCL27, GPNMB, FGF23, IL32, PSME1 or TBCB; preferably, the protein markers further include any one or more of the following: CX3CL1, CLEC10A, IL10, NTRK3, SNCG, PMVK, PTPN; preferably, the protein markers include: (1) CCL19, CX3CL1, IL10, CPA2 and DPEP1; or (2) CCL19, CX3CL1, IL.17C, CPA2 and DPEP1; or (3) CCL19, IL10, CPA2, NTRK3 and DPEP1; or (4) CCL19, CX3CL1, CPA2 and DPEP1; or (5) CX3CL1, IL10, CPA2 and DPEP1; or (6) CX3CL1, CXCL6, IL.17C, CPA2 and PMVK; or (7) CX3CL1, CXCL6, IL.17C, CPA2 and PTPN1; or (8) CX3CL1, CXCL6, CPA2, GPNMB and PTPN1; or (9) CX3CL1, CXCL6, CPA2, GPNMB and PMVK; or (10) CX3CL1, CXCL6, IL.17C, CPA2 and TBCB; or (11) CLEC10A, CPA2, NTRK3 and SNCG; or (12) CLEC10A, CPA2, NTRK3 and DPEP1; or (13) CLEC10A, CPA2, NTRK3, DPEP1 and SNCG; or (14) IL10, CLEC10A, CPA2, NTRK3 and SNCG; or (15) CLEC10A, NTRK3 and DPEP1; or (16) CX3CL1, CXCL6, IL10, CLEC10A and DPEP1; or (17) CX3CL1, CXCL6, IL.17 C, CLEC10A and DPEP1; or (18) CX3CL1, CXCL6, CLEC10A and DPEP1; or (19) CX3CL1, CXCL6, CLEC10A, DPEP1 and PSME1; or (20) CX3CL1, CXCL6, FGF.23, CLEC10A and DPEP1; or (21) IL10, CCL27 and DPEP1; or (22) IL10, CPA2 and DPEP1; or (23) IL10, CCL27, DPEP1 and IL32; or (24) IL10, CLEC10A, CPA2 and DPEP1; or (25) IL10, CLEC10A and DPEP1; or (26) CPA2 and DPEP1; or (27) CPA2, DPEP1 and CX3CL1; or (28) CPA2, DPEP1, CX3CL1 and CLEC10A; or (29) CPA2, DPEP1, CX3CL1, CLEC10A and IL10; preferably, the prediction result according to the model output is that the predict function is used with type='response', and the output probability value is ≥0.5, indicating that the subject has Parkinson's disease or is at high risk, and the probability value is <0.5, indicating that the subject does not have Parkinson's disease or is at low risk.

[0015] To achieve the above-mentioned purpose, according to a fourth aspect of the present application, a screening method for diagnosing or predicting a protein marker of Parkinson's disease is provided, which comprises: obtaining the Parkinson's disease condition and biological samples of a sample population, detecting the biological samples to obtain the protein content in the biological samples, the sample population comprising healthy people and Parkinson's patients; screening significant differential proteins that are different between Parkinson's patients and healthy people according to the protein content by a statistical method; constructing a prediction model by a machine learning training method using the content of the significant differential proteins of the sample population and the Parkinson's disease condition; and the protein appearing in the prediction model more than a threshold value is the protein marker.

[0016] Further, the random splitting is performed multiple times, the prediction model is constructed multiple times, and the prediction model is cross-validated multiple times; preferably, the statistical method includes any one or more of the following: T test, generalized linear model, rank test, logistic regression, fold change, multiple hypothesis testing correction, or Kruskal-Wallis test; preferably, the biological sample includes plasma; preferably, constructing the prediction model includes: dividing the sample population into a training set and a test set, establishing a mapping relationship between the content of the significantly different proteins and the Parkinson's disease condition of the training set or the test set, and obtaining a training data set or a test data set, respectively; using a machine learning training method and adopting five-fold cross-validation, a prediction model is constructed; preferably, the machine learning training includes any one or more of the following: logistic regression model, support vector machine model, or naive Bayes model; preferably, during the construction of the prediction model, an analysis method is used to analyze the modeling effect; preferably, the analysis method includes one or more of the following: lasso regression, Bayesian information criterion, or stepwise regression.

[0017] In order to achieve the above-mentioned purpose, according to the fifth aspect of the present application, a method for constructing a diagnosis model is provided, which uses the protein markers obtained by the above-mentioned screening method and applies a machine learning training method to construct a diagnosis model.

[0018] Further, the protein markers comprise CPA2 and / or DPEP1; preferably, the protein markers further comprise any one or more of: CXCL6, IL17C, CCL19, CCL27, GPNMB, FGF23, IL32, PSME1, or TBCB; preferably, the protein markers further comprise any one or more of: CX3CL1, CLEC10A, IL10, NTRK3, SNCG, PMVK, PTPN; preferably, the protein markers comprise: (1) CCL19, CX3CL1, IL10, CPA2, and DPEP1; or (2) CCL19, CX3CL1, IL.17C, CPA2, and DPEP1; or (3) CCL19, IL10, CPA2, NTRK3, and DPEP1; or (4) CCL19, CX3CL1, CPA2, and DPEP1; or (5) CX3CL1, IL10, CPA2, and DPEP1; or (6) CX3CL1, CXCL6, IL.17C, CPA2, and PMVK; or (7) CX3CL1, CXCL6, IL.17C, CPA2, and PTPN1; or (8) CX3CL1, CXCL6, CPA2, GPNMB, and PTPN1; or (9) CX3CL1, CXCL6, CPA2, GPNMB, and PMVK; or (10) CX3CL1, CXCL6, IL.17C, CPA2, and TBCB; or (11) CLEC10A, CPA2, NTRK3, and SNCG; or (12) CLEC10A, CPA2, NTRK3, and DPEP1; or (13) CLEC10A, CPA2, NTRK3, DPEP1, and SNCG; or (14) IL10, CLEC10A, CPA2, NTRK3, and SNCG; or (15) CLEC10A, NTRK3, and DPEP1; or (16) CX3CL1, CXCL6, IL10, CLEC10A, and DPEP1; or (17) CX3CL1, CXCL6, IL.17C, CLEC10A, and DPEP1; or (18) CX3CL1, CXCL6, CLEC10A, and DPEP1; or (19) CX3CL1, CXCL6, CLEC10A, DPEP1, and PSME1; or (20) CX3CL1, CXCL6, FGF.23. CLEC10A and DPEP1; or (21) IL10, CCL27 and DPEP1; or (22) IL10, CPA2 and DPEP1; or (23) IL10, CCL27, DPEP1 and IL32; or (24) IL10, CLEC10A, CPA2 and DPEP1; or (25) IL10, CLEC10A and DPEP1; or (26) CPA2 and DPEP1; or (27) CPA2, DPEP1 and CX3CL1; or (28) CPA2, DPEP1, CX3CL1 and CLEC10A; or (29) CPA2, DPEP1, CX3CL1, CLEC10A and IL10.

[0019] To achieve the above object, according to a sixth aspect of the present application, a computer readable storage medium is provided, which comprises a stored program, wherein when the program is run, the device where the storage medium is located is controlled to execute the screening method, the construction method or the diagnostic model.

[0020] To achieve the above object, according to a seventh aspect of the present application, a processor is provided, which is used to run a program, wherein when the program is run, the screening method, the construction method or the diagnostic model is executed.

[0021] To achieve the above object, according to an eighth aspect of the present application, a device for screening protein markers for diagnosing or predicting Parkinson's disease is provided, which comprises: an acquisition detection module configured to acquire the Parkinson's disease condition of a sample population and a biological sample, detect the biological sample to obtain the protein content in the biological sample, and the sample population comprises healthy people and Parkinson's patients; a difference statistics module configured to screen significant difference proteins that have differences between Parkinson's patients and healthy people by statistical methods according to the protein content; a model construction module configured to construct a plurality of prediction models by using the content of the significant difference proteins of the sample population and the Parkinson's disease condition; and a marker determination module configured to determine that a protein appearing more than a threshold value in the plurality of prediction models is a protein marker.

[0022] Further, the proteins to be detected in the biological sample include CPA2 and / or DPEP1; preferably, the proteins to be detected further include any one or more of the following: CXCL6, IL17C, CCL19, CCL27, GPNMB, FGF23, IL32, PSME1 or TBCB; preferably, the proteins to be detected further include any one or more of the following: CX3CL1, CLEC10A, IL10, NTRK3, SNCG, PMVK, PTPN.

[0023] By applying the technical scheme of the present application, a plurality of protein markers capable of diagnosing or predicting early Parkinson's disease are screened, and by using one or more of the protein markers, a product for diagnosing Parkinson's disease or predicting Parkinson's disease can be prepared, thereby facilitating the diagnosis or prediction of the condition or risk of Parkinson's disease of a subject and improving the effect of diagnosis or prediction. BRIEF DESCRIPTION OF DRAWINGS

[0024] The accompanying drawings, which form a part of the present application, are intended to provide further understanding of the present application, and are used to interpret the illustrative embodiments of the present application and their descriptions, and do not constitute improper limitations on the present application. In the drawings:

[0025] Figure 1 A ROC curve diagram of a prediction model composed of 18 protein markers according to Embodiment 1 of the present application is shown.

[0026] Figure 2 A ROC curve diagram of a prediction model composed of 5 protein markers according to Embodiment 1 of the present application is shown. DETAILED DESCRIPTION

[0027] It should be noted that the embodiments in the present application and the features in the embodiments can be combined with each other without conflict. The present application will be described in detail below with reference to the embodiments.

[0028] As mentioned in the background, the prior art is difficult to meet the needs of the diagnosis of Parkinson's disease, especially early Parkinson's disease. The present application provides biomarkers for identifying Parkinson's disease from the perspective of protein molecules. By obtaining the content of the protein markers, a prediction probability is given by model calculation, which can assist in the diagnosis of Parkinson's disease and provide more definite and effective diagnosis results for the subject.

[0029] Therefore, in the present application, the inventors attempt to screen the proteins of the subject, find protein markers related to Parkinson's disease, use one or more protein markers to prepare a product for diagnosing Parkinson's disease or predicting Parkinson's disease, and predict the Parkinson's disease condition or risk of Parkinson's disease of the subject, and thus a series of protection schemes of the present application are proposed.

[0030] In a first typical embodiment of the present application, an application of a protein marker in the preparation of a product for diagnosing Parkinson's disease or predicting Parkinson's disease is provided, and the protein marker comprises CPA2 and / or DPEP1.

[0031] In a preferred embodiment, the protein markers further comprise any one or more of: CXCL6, IL17C, CCL19, CCL27, GPNMB, FGF23, IL32, PSME1, or TBCB; preferably the protein markers further comprise any one or more of CX3CL1, CLEC10A, IL10, NTRK3, SNCG, PMVK, PTPN.

[0032] In a preferred embodiment, the protein markers are selected from any one of the following groups: (1) CCL19, CX3CL1, IL10, CPA2, and DPEP1; or (2) CCL19, CX3CL1, IL.17C, CPA2, and DPEP1; or (3) CCL19, IL10, CPA2, NTRK3, and DPEP1; or (4) CCL19, CX3CL1, CPA2, and DPEP1; or (5) CX3CL1, IL10, CPA2, and DPEP1; or (6) CX3CL1, CXCL6, IL.17C, CPA2, and PMVK; or (7) CX3CL1, CXCL6, IL.17C, CPA2, and PTPN1; or (8) CX3CL1, CXCL6, CPA2, GPNMB, and PTPN1; or (9) CX3CL1, CXCL6, CPA2, GPNMB, and PMVK; or (10) CX3CL1, CXCL6, IL.17C, CPA2, and TBCB; or (11) CLEC10A, CPA2, NTRK3, and SNCG; or (12) CLEC10A, CPA2, NTRK3, and DPEP1; or (13) CLEC10A, CPA2, NTRK3, DPEP1, and SNCG; or (14) IL10, CLEC10A, CPA2, NTRK3, and SNCG; or (15) CLEC10A, NTRK3, and DPEP1; or (16) CX3CL1, CXCL6, IL10, CLEC10A, and DPEP1; or (17) CX3CL1, CXCL6, IL.17C, CLEC10A, and DPEP1; or (18) CX3CL1, CXCL6, CLEC10A, and DPEP1; or (19) CX3CL1, CXCL6, CLEC10A, DPEP1, and PSME1; or (20) CX3CL1, CXCL6, FGF.23, CLEC10A, and DPEP1; or (21) IL10, CCL27, and DPEP1; or (22) IL10, CPA2, and DPEP1; or (23) IL10, CCL27, DPEP1, and IL32; or (24) IL10, CLEC10A, CPA2, and DPEP1; or (25) IL10, CLEC10A, and DPEP1; or (26) CPA2, and DPEP1; or (27) CPA2, DPEP1, and CX3CL1; or (28) CPA2, DPEP1, CX3CL1, and CLEC10A; or (29) CPA2, DPEP1, CX3CL1, CLEC10A, and IL10.

[0033] Proteomics is a science that studies the composition and variation of the proteome of a cell, tissue or organism. After genomics, proteomics research has become a new frontier for understanding complex biological systems. Studies have shown that proteomics has great potential in identifying disease-related biomarkers, monitoring disease progression, and evaluating treatment response. In addition, due to the role of proteins as target molecules, it also plays an important role in drug development.

[0034] The above protein markers and combinations of protein markers have the ability to prepare products for diagnosing Parkinson's disease or predicting Parkinson's disease. Using the above products, it can be determined whether a subject has Parkinson's disease or predicts the risk of Parkinson's disease. Such determination and prediction ability and effect cannot be predicted before the implementation of this embodiment. In this application, Parkinson's disease and Parkinson's disease refer to the neurodegenerative disease Parkinson's disease (PD).

[0035] In a preferred embodiment, the application comprises detecting the content of the protein marker by using a detection reagent; preferably, the detection method of the content of the protein marker includes but is not limited to Olink technology, western blot or enzyme-linked immunosorbent assay (ELISA); preferably, the detection reagent is an antibody of the protein marker; preferably, the antibody is arranged on a solid support; preferably, the solid support includes an enzyme-labeled plate, a membrane carrier or a microsphere; preferably, the membrane carrier is selected from nitrocellulose membrane, glass cellulose membrane or nylon membrane; preferably, the antibody is a monoclonal antibody or a polyclonal antibody; preferably, the related product includes a kit or an electronic prediction device.

[0036] All clinically detected protein methods are suitable for this application, and the corresponding detection reagent is also selected according to different detection methods. The antibody is preferably arranged in a pre-coated form. Preferably, the pre-coated antibody is coated on a solid support; the specific solid support is reasonably designed according to the needs. More preferably, the solid support includes an enzyme-labeled plate (mostly polystyrene material), a membrane carrier or a microsphere; further preferably, the membrane carrier includes a nitrocellulose membrane (the most widely used), a glass cellulose membrane or a nylon membrane, and more preferably, the membrane carrier is also coated with a positive control, and the corresponding protein marker and positive control are arranged in sequence on the nitrocellulose membrane according to the detection order.

[0037] In the above applications, the content of the protein markers can be detected by using the detection reagent, and the content detected can be used to judge or predict Parkinson's disease. By using proteomics such as Olink technology, or existing technologies such as western blot and ELISA, and using common detection reagents such as antibodies, the content of multiple protein markers in a biological sample can be detected, the sample amount required is small, the accuracy is high, the content of multiple protein markers in a biological sample can be detected, and further diagnosis or prediction of Parkinson's disease can be performed. The detection reagent used in the above detection method is a conventional detection reagent in the corresponding method. The related products for diagnosing or predicting Parkinson's disease include but are not limited to a kit or an electronic prediction device.

[0038] In a second typical embodiment of the present application, a kit for diagnosing or predicting Parkinson's disease in a subject is provided, which comprises a detection reagent for detecting the content of protein markers, and the protein markers comprise CPA2 and / or DPEP1.

[0039] In a preferred embodiment, the protein markers further comprise any one or more of: CXCL6, IL17C, CCL19, CCL27, GPNMB, FGF23, IL32, PSME1, or TBCB; preferably the protein markers further comprise any one or more of: CX3CL1, CLEC10A, IL10, NTRK3, SNCG, PMVK, PTPN; preferably the protein markers comprise: (1) CCL19, CX3CL1, IL10, CPA2, and DPEP1; or (2) CCL19, CX3CL1, IL.17C, CPA2, and DPEP1; or (3) CCL19, IL10, CPA2, NTRK3, and DPEP1; or (4) CCL19, CX3CL1, CPA2, and DPEP1; or (5) CX3CL1, IL10, CPA2, and DPEP1; or (6) CX3CL1, CXCL6, IL.17C, CPA2, and PMVK; or (7) CX3CL1, CXCL6, IL.17C, CPA2, and PTPN1; or (8) CX3CL1, CXCL6, CPA2, GPNMB, and PTPN1; or (9) CX3CL1, CXCL6, CPA2, GPNMB, and PMVK; or (10) CX3CL1, CXCL6, IL.17C, CPA2, and TBCB; or (11) CLEC10A, CPA2, NTRK3, and SNCG; or (12) CLEC10A, CPA2, NTRK3, and DPEP1; or (13) CLEC10A, CPA2, NTRK3, DPEP1, and SNCG; or (14) IL10, CLEC10A, CPA2, NTRK3, and SNCG; or (15) CLEC10A, NTRK3, and DPEP1; or (16) CX3CL1, CXCL6, IL10, CLEC10A, and DPEP1; or (17) CX3CL1, CXCL6, IL.17C, CLEC10A, and DPEP1; or (18) CX3CL1, CXCL6, CLEC10A, and DPEP1; or (19) CX3CL1, CXCL6, CLEC10A, DPEP1, and PSME1; or (20) CX3CL1, CXCL6, FGF.23, CLEC10A and DPEP1; or (21) IL10, CCL27 and DPEP1; or (22) IL10, CPA2 and DPEP1; or (23) IL10, CCL27, DPEP1 and IL32; or (24) IL10, CLEC10A, CPA2 and DPEP1; or (25) IL10, CLEC10A and DPEP1; or (26) CPA2 and DPEP1; or (27) CPA2, DPEP1 and CX3CL1; or (28) CPA2, DPEP1, CX3CL1 and CLEC10A; or (29) CPA2, DPEP1, CX3CL1 and CLEC10A, IL10.

[0040] The kit comprises detection reagents for detecting the content of one or more protein markers described above. Using the detection reagents and corresponding detection methods, detection equipment, the content of the protein markers can be detected. Preferably, the detection reagents comprise detection reagents required for detecting the protein markers using existing technologies such as Olink technology.

[0041] In a third typical embodiment of the present application, an electronic device for diagnosing or predicting Parkinson's disease is provided, comprising: an acquisition detection module configured to acquire a biological sample, detect the biological sample, and obtain the content of protein markers in the biological sample; a diagnosis or prediction module, wherein the diagnosis or prediction module is configured to establish a regression equation using a logistic regression model, a support vector machine model or a naive Bayes model and the content of the protein markers, construct a model, and then output a prediction result according to the model; in a preferred embodiment, the biological sample is a plasma sample; preferably, the protein markers in the biological sample comprise CPA2 and / or DPEP1; preferably, the protein markers in the biological sample further comprise any one or more of the following: CXCL6, IL17C, CCL19, CCL27, GPNMB, FGF23, IL32, PSME1 or TBCB; preferably, the protein markers further comprise any one or more of the following: CX3CL1, CLEC10A, IL10, NTRK3, SNCG, PMVK, PTPN; preferably, the protein markers comprise: (1) CCL19, CX3CL1, IL10, CPA2 and DPEP1; or (2) CCL19, CX3CL1, IL.17C, CPA2 and DPEP1; or (3) CCL19, IL10, CPA2, NTRK3 and DPEP1; or (4) CCL19, CX3CL1, CPA2 and DPEP1; or (5) CX3CL1, IL10, CPA2 and DPEP1; or (6) CX3CL1, CXCL6, IL.17C, CPA2 and PMVK; or (7) CX3CL1, CXCL6, IL.17C, CPA2 and PTPN1; or (8) CX3CL1, CXCL6, CPA2, GPNMB and PTPN1; or (9) CX3CL1, CXCL6, CPA2, GPNMB and PMVK; or (10) CX3CL1, CXCL6, IL.17C, CPA2 and TBCB; or (11) CLEC10A, CPA2, NTRK3 and SNCG; or (12) CLEC10A, CPA2, NTRK3 and DPEP1; or (13) CLEC10A, CPA2, NTRK3, DPEP1 and SNCG; or (14) IL10, CLEC10A, CPA2, NTRK3 and SNCG; or (15) CLEC10A, NTRK3 and DPEP1; or (16) CX3CL1, CXCL6, IL10, CLEC10A and DPEP1; or (17) CX3CL1, CXCL6, IL.17 C, CLEC10A and DPEP1; or (18) CX3CL1, CXCL6, CLEC10A and DPEP1; or (19) CX3CL1, CXCL6, CLEC10A, DPEP1 and PSME1; or (20) CX3CL1, CXCL6, FGF.23, CLEC10A and DPEP1; or (21) IL10, CCL27 and DPEP1; or (22) IL10, CPA2 and DPEP1; or (23) IL10, CCL27, DPEP1 and IL32; or (24) IL10, CLEC10A, CPA2 and DPEP1; or (25) IL10, CLEC10A and DPEP1; or (26) CPA2 and DPEP1; or (27) CPA2, DPEP1 and CX3CL1; or (28) CPA2, DPEP1, CX3CL1 and CLEC10A; or (29) CPA2, DPEP1, CX3CL1, CLEC10A and IL10; preferably, the prediction result according to the model output is that the predict function is used with type='response', and the output probability value is ≥0.5, indicating that the subject has Parkinson's disease or is at high risk, and the probability value is <0.5, indicating that the subject does not have Parkinson's disease or is at low risk.

[0042] In a fourth typical embodiment of the present application, a screening method for diagnosing Parkinson's disease or predicting protein markers of Parkinson's disease is provided, which comprises: obtaining the Parkinson's disease status and biological samples of a sample population, detecting the biological samples to obtain the protein content in the biological samples, the sample population including healthy people and Parkinson's patients; screening significant differential proteins that are different between Parkinson's patients and healthy people according to the protein content through a statistical method; using the content of the significant differential proteins of the sample population and the Parkinson's disease status, constructing a prediction model through a machine learning training method; and the proteins appearing more than a threshold value in the prediction model are protein markers.

[0043] In the screening method of the above protein marker, first, biological samples of a sample population including healthy people and Parkinson's patients and Parkinson's disease conditions are obtained. The biological samples are subjected to protein detection, a certain number of proteins are selected as detection, the detection of the detection proteins in the biological samples is performed, and the content of the detection proteins in each biological sample is obtained. The proteins that are different between Parkinson's disease patients and healthy people are tested using a statistical method, which are significant difference proteins. The significant difference proteins indicate that the amount of these proteins in the sample population is related to the disease condition, and the amount of significant difference proteins in healthy people and Parkinson's patients is significantly different, but it cannot be confirmed that the significant difference proteins have specific effects on diagnosis or prediction, and it is not clear how to diagnose or predict. In order to ensure and improve the effect of diagnosing or predicting Parkinson's disease, the content of the significant difference proteins of the sample population and the disease condition are modeled using machine learning and other methods, and a prediction model for diagnosing or predicting Parkinson's disease is constructed. The proteins appearing in the prediction model are all proteins selected from the significant difference proteins, which are higher than the threshold value and have good prediction effect, that is, protein markers.

[0044] In a preferred embodiment, the number of random splits is multiple, the number of constructed prediction models is multiple, and the prediction models are cross-validated multiple times; preferably, the statistical method includes any one or more of the following: T-test, generalized linear model, rank test, logistic regression, fold change, multiple hypothesis testing correction, or Kruskal-Wallis test; preferably, the biological samples include plasma; preferably, constructing the prediction model includes: dividing the sample population into a training set and a test set, establishing a mapping relationship between the content of the significant difference proteins of the training set or the test set and the Parkinson's disease conditions, and obtaining a training data set or a test data set, respectively; using the training data set for machine learning training to construct the prediction model; using the machine learning training method and adopting five-fold cross-validation to construct the prediction model; preferably, the machine learning training includes any one or more of the following: logistic regression model, support vector machine model, or naive Bayes model; preferably, during the construction of the prediction model, an analysis method is used to analyze the modeling effect; preferably, the analysis method includes one or more of LASSO regression, BIC (Bayesian Information Criterion), and stepwise regression.

[0045] The statistical methods for determining the significantly different proteins include, but are not limited to, one or more of T-test, Wilcoxontest, logistic regression, KW test, Fold Change, Bonferroni correction, or other existing statistical methods. When constructing the prediction model, the disease conditions and the contents of the significantly different proteins of each sample are summarized to obtain the summarized data of different individuals, machine learning training is performed, and five-fold cross-validation is used to construct the prediction model.

[0046] The machine learning training includes, but is not limited to, one or more of logistic regression model, SVM (support vector machines), NBM (Naive Bayes Model). Using the above different machine learning methods, different prediction models can be constructed. In the process of five-fold cross-validation, the verification results of different prediction models are compared, and the model with the best prediction effect can be obtained. In the process of establishing the prediction model, one or more of the analysis methods including, but not limited to, LASSO regression, BIC (Bayesian Information Criterion), and stepwise regression are used to analyze the established prediction model, and the prediction model with high effectiveness is selected.

[0047] In the fifth typical embodiment of the present application, a method for constructing a diagnosis model is provided. The protein markers obtained by the above screening method are used to construct a diagnosis model by using a machine learning training method, which includes logistic regression, support vector machines, or Naive Bayes Model, preferably logistic regression model.

[0048] In a preferred embodiment, the protein markers comprise CPA2 and / or DPEP1; preferably the protein markers further comprise any one or more of: CXCL6, IL17C, CCL19, CCL27, GPNMB, FGF23, IL32, PSME1, or TBCB; preferably the protein markers further comprise any one or more of: CX3CL1, CLEC10A, IL10, NTRK3, SNCG, PMVK, PTPN1; preferably the protein markers comprise: (1) CCL19, CX3CL1, IL10, CPA2, and DPEP1; or (2) CCL19, CX3CL1, IL.17C, CPA2, and DPEP1; or (3) CCL19, IL10, CPA2, NTRK3, and DPEP1; or (4) CCL19, CX3CL1, CPA2, and DPEP1; or (5) CX3CL1, IL10, CPA2, and DPEP1; or (6) CX3CL1, CXCL6, IL.17C, CPA2, and PMVK; or (7) CX3CL1, CXCL6, IL.17C, CPA2, and PTPN1; or (8) CX3CL1, CXCL6, CPA2, GPNMB, and PTPN1; or (9) CX3CL1, CXCL6, CPA2, GPNMB, and PMVK; or (10) CX3CL1, CXCL6, IL.17C, CPA2, and TBCB; or (11) CLEC10A, CPA2, NTRK3, and SNCG; or (12) CLEC10A, CPA2, NTRK3, and DPEP1; or (13) CLEC10A, CPA2, NTRK3, DPEP1, and SNCG; or (14) IL10, CLEC10A, CPA2, NTRK3, and SNCG; or (15) CLEC10A, NTRK3, and DPEP1; or (16) CX3CL1, CXCL6, IL10, CLEC10A, and DPEP1; or (17) CX3CL1, CXCL6, IL.17C, CLEC10A, and DPEP1; or (18) CX3CL1, CXCL6, CLEC10A, and DPEP1; or (19) CX3CL1, CXCL6, CLEC10A, DPEP1, and PSME1; or (20) CX3CL1, CXCL6, FGF.23. CLEC10A and DPEP1; or (21) IL10, CCL27 and DPEP1; or (22) IL10, CPA2 and DPEP1; or (23) IL10, CCL27, DPEP1 and IL32; or (24) IL10, CLEC10A, CPA2 and DPEP1; or (25) IL10, CLEC10A and DPEP1; or (26) CPA2 and DPEP1; or (27) CPA2, DPEP1 and CX3CL1; or (28) CPA2, DPEP1, CX3CL1 and CLEC10A; or (29) CPA2, DPEP1, CX3CL1, CLEC10A and IL10.

[0049] In a sixth typical implementation of the present application, a computer readable storage medium is provided, the storage medium comprising a stored program, wherein the program, when executed, controls a device on which the storage medium is located to perform the screening method, the constructing method or the constructed model.

[0050] In a seventh typical implementation of the present application, a processor is provided, the processor being configured to execute a program, wherein the program, when executed, performs the screening method, the constructing method or the constructed model.

[0051] The present application can be used in a variety of general purpose or special purpose computing system environments or configurations. Examples of computing systems, environments, and / or configurations that can be used include personal computer systems, server computer systems, handheld or laptop device computing systems, mobile computing devices, multiprocessor systems, microprocessor-based systems, set top boxes, programmable consumer electronics, network PC systems, minicomputer systems, mainframe computer systems, distributed computing environments that include any of the above systems or devices, and the like.

[0052] Obviously, those skilled in the art should understand that some of the modules or steps of the present application described above can be realized by general computing devices, and they can be centralized on a single computing device or distributed on a network composed of multiple computing devices. Alternatively, they can be realized by program codes executable by computing devices, so that they can be stored in storage devices and executed by computing devices, or they can be respectively manufactured into individual integrated circuit modules, or multiple modules or steps among them can be manufactured into a single integrated circuit module. Thus, the present application is not limited to any specific combination of hardware and software.

[0053] In an eighth typical implementation of the present application, there is provided a device for screening protein markers for diagnosing or predicting Parkinson's disease, comprising: an acquisition detection module configured to acquire the Parkinson's disease conditions of a sample population and biological samples, detect the biological samples to obtain the protein content in the biological samples, the sample population comprising healthy people and Parkinson's patients; a difference statistics module configured to screen significant difference proteins that are different between the Parkinson's patients and the healthy people according to the protein content through a statistical method; a model construction module configured to construct a plurality of prediction models using the content of the significant difference proteins of the sample population and the Parkinson's disease conditions; and a marker determination module configured to determine that a protein appearing more than a threshold value in the plurality of prediction models is a protein marker.

[0054] In a preferred embodiment, the proteins to be detected in the biological samples include CPA2 and / or DPEP1; preferably, the proteins to be detected further include any one or more of: CXCL6, IL17C, CCL19, CCL27, GPNMB, FGF23, IL32, PSME1, or TBCB; preferably, the proteins to be detected further include any one or more of: CX3CL1, CLEC10A, IL10, NTRK3, SNCG, PMVK, PTPN.

[0055] By using the above device, the protein markers for diagnosing or predicting Parkinson's disease can be screened by detecting the protein content of the above proteins to be detected in the biological samples of the sample population, which facilitates the screening, updating, or correction of the protein markers when the prediction effect of the protein markers is affected by factors such as the region, environment, and the like of the subject.

[0056] The beneficial effects of the present application will be further explained in detail below with reference to specific examples.

[0057] Example 1:

[0058] 1. Materials and methods

[0059] 1.1 Clinical sample inclusion

[0060] The research subjects of the present project were 88 Parkinson's disease patients, and the control group was 88 healthy people, a total of 176 samples, and all the patients received blood sampling. The age range of the patients was 28-85 years old, including 85 male samples and 89 female samples.

[0061] All subjects obtained individual basic information, including age, gender, height, BMI, smoking, alcohol consumption, constipation, antibiotic use, and the like, through a self-made health questionnaire, as shown in Table 1.

[0062] The above research plan conforms to the ethical standards of human experiments and is approved by the ethics committee. The subjects are informed and consent before the test and sign the informed consent form.

[0063] Table 1 Clinical data of all 176 samples

[0064]

[0065]

[0066]

[0067]

[0068]

[0069]

[0070] Note: " / " indicates that the data does not exist. "NA" indicates that the data is missing. BMI is Body Mass Index (BMI) HY score indicates Parkinson's disease HY classification, 1 = single disease; 1.5 = single disease combined with trunk (axis) symptoms; 2 = bilateral disease but no balance disorder; 2.5 = mild bilateral disease, can recover from the pull test; 3 = mild to moderate bilateral disease, cannot recover from the pull test, but can take care of themselves; 4 = severe disability, but can stand or walk on their own, 5 = wheelchair or bedridden, completely dependent on others.

[0071] 1.2 Olink protein detection platform

[0072] Protein sequencing uses Olink technology from a Swedish innovative proteomics detection program company named Olink Proteomics, which uses a unique proteomics detection technology called proximity extension assay (PEA). PEA technology uses a pair of matched antibodies, respectively coupled with unique DNA oligonucleotide chains, to form a probe with high specificity, which binds to the detected substance, and the two similar DNA chains are complementary and extended to generate a new DNA molecule template; then through qPCR amplification, the signal is amplified for detection. PCR amplification enriches the library template, and sodium hydroxide denaturation produces single-stranded DNA fragments. Sequencing and bioinformatics analysis are completed by Shenzhen Carbon Cloud Intelligence.

[0073] 1.3 Sample quality control

[0074] The platform uses proximity extension assay (PEA) technology to detect and analyze 274 protein indicators in 176 plasma samples of the above-mentioned wide Chinese medicine project, a total of three panels. By adding four internal quality control samples in each sample, the overall detection quality and the detection quality of each sample are monitored.

[0075] A total of three chips are detected, and two sample versions are used for each chip, so a total of 6 times are detected, as shown in Table 2.1.

[0076] Table 2.1 Combination directory

[0077]

[0078] The overall quality control is evaluated according to the standard deviation of the internal quality control of the whole plate. The standard deviation is less than 0.2, which can pass the quality control. The 6 detections of this project pass the overall quality control. The specific is shown in Table 2.2.

[0079] The single-hole sample quality control is evaluated by calculating the deviation of the value of the internal quality control of each sample from the median of the overall quality control. The sample whose deviation from the median is not more than 0.3 can pass the quality control. In this test of the project, the overall sample quality control pass rate is 97%, which can pass the quality control (the pass rate is an integer, and after rounding off, it reaches 97%). The sample quality control results are shown in Table 2.3.

[0080] Table 2.2 Standard deviation of each plate of overall quality control

[0081]

[0082] Table 2.3 Sample quality control pass rate (%)

[0083]

[0084] Considering the Inflammation, Neuro Exploratory, and Neurology 3 data sets, selecting samples that are qualified in all three data sets, the final qualified samples are 170, of which 86 are in the disease group and 84 are in the healthy group.

[0085] 2. Screening of significantly different proteins

[0086] In order to compare the protein substances of the control group and the disease group, the significantly different proteins that exhibit differences between the above two groups are screened. The present application uses R software to analyze the different protein substances, and uses statistical methods to analyze the 274 proteins of the 170 samples by combining significance statistical test (generalized linear model or T test, etc.) and difference multiple (Fold change). When the protein significance statistical test is P<0.05, it is considered as a significantly different protein. A total of 31 significantly different proteins are obtained, as shown in Table 3.

[0087] Table 3. 31 significantly different proteins between Parkinson’s disease patients and healthy people

[0088]

[0089]

[0090] Note: UniPort ID is the number of this protein in UniPort database (www.uniport.org).

[0091] 2.1 Screening disease prediction protein markers

[0092] To evaluate the proteomics of Parkinson’s disease, i.e., the diagnostic value of the above 31 significantly different proteins, the applicant constructed a logistic regression model that can detect Parkinson’s patients. The above 170 samples were divided into a training set of 70% and a test set of 30%, where the training set included 57 samples of Parkinson’s disease patients and 66 samples of healthy people, and the test set included 30 samples of Parkinson’s disease patients and 17 samples of healthy people. The mapping relationship between the content of the significantly different proteins in the training set or test set and the Parkinson’s disease condition was obtained to obtain the training data set or test data set, respectively. The content of the significantly different proteins and the Parkinson’s disease condition (i.e., the training data set) were used to train the logistic regression model, and the BIC analysis method was used to find the optimal model to determine the protein marker that had the greatest impact on the disease (R.4.0.5, BMA3.18.15 software package) (Adrian Raftery, Jennifer Hoeting, Chris Volinsky, Ian Painter and Ka Yee Yeung (2021)), thereby selecting the optimal set of protein markers. The model was tested on the test set, and the prediction error was determined.

[0093] ① Regarding the regression model, by using the “BMA3.18.1 package” in the R software (R vision 4.0.5) version 4.0.5, the training data set was input, including the content of the 31 significantly different proteins in the training set and the Parkinson’s disease condition (the disease condition is a vector, 1 represents the control, and 0 represents the case).

[0094] BMA package in R software to build classification models, and the function was used to screen the best disease prediction model. The predict() function was used to predict the classification effect of the model in the test set. The bic.glm function outputs the prediction results (variable effect probability on disease (p! = 0)); estimated model bias (EV), standard deviation (SD), five potential best models (model 1 - model 5). As shown in Table 4, the larger the posterior probability and the smaller the BIC, the better the model. Table 4 shows that the first list represents the protein category, and the numbers under the model 1, model 2, model 3, model 4, and model 5 columns indicate that the model includes the protein. For example: the proteins included in model 1 are CX3CL1, CXCL6, IL 17C, CPA2, and PMVK.

[0095] Table 4. The top five prediction combinations with the largest BIC posterior probability

[0096]

[0097]

[0098] ③Five repeated cross-validation (five-fold cross-validation) was performed on the model.

[0099] Based on the logistic regression model, Bayesian model averaging (BMC) was used for five repeated validation. In each cross-validation, 170 samples were randomly divided into 70% training set and 30% test set based on the fixed ratio of diseased group to healthy group, so the training set of five validations was 123 samples (including 57 samples of Parkinson's disease patients and 66 samples of healthy people). Bayesian model averaging averages the best model in the model class according to the approximate posterior probability (posterior probability - postprob). All combination groups of the top five protein markers with the largest posterior probability were listed each time, and 25 prediction model combinations were obtained, as shown in Table 5. At the same time, the proteins were sorted according to the number of occurrences, and the sorting of the 18 proteins with the highest number of occurrences is shown in Table 6, in which the most frequently occurring proteins are DPEP1, CPA2, CX3CL1, CLEC10A, and IL10.

[0100] Table 5. 25 prediction model combinations

[0101]

[0102]

[0103] Table 4 shows 5 models obtained by one cross-validation; Table 5 shows 25 models (including the 5 models in Table 4) obtained by 5 repetitions.

[0104] (4) Calculate the discrimination ability of various combinations of prediction models

[0105] The discrimination ability of the above 25 prediction models was calculated, and the detection ability of the validation random significant protein combinations was attempted in multiple ways, such as using logistic regression models, NBM models, and SVC models, and it was found that the prediction results were good. The following is an example of the effect of the logistic regression model and BIC (Bayesian Information Criterion method). In the following examples, different combinations of protein markers were used to build logistic regression models, the training data set and BIC were used to train and screen the models, and the test data set was used to verify the models, calculate AUC and other numerical values, as shown in the last 3 columns of Table 5, and the test error was calculated. The prediction model was used to calculate the probability of Parkinson's disease, and the ROC (R 4.0.5, pROC 1.18.0 package) was drawn.

[0106] (5) Screening of protein markers

[0107] The above models 1 to 25 appeared 18 kinds of significant differential proteins, i.e. 18 kinds of protein markers used to combine prediction models, as shown in Table 6. The more times a protein appears in each model, the more it indicates that the protein is a key role in building a model.

[0108] Table 6. 18 protein markers

[0109]

[0110]

[0111] Our research results show that the above 18 proteins perform well in the training set, as shown in Table 7, the area under the ROC curve (AUC) is 90.89%. Figure 1

[0112] 2.2 Construction of diagnostic models using randomly combined protein markers screened

[0113] The higher the frequency in the screening process in 2.1, the stronger the correlation with Parkinson's disease diagnosis, so the top 5 proteins (IL10, CX3CL1, DPEP1, CLEC10A, and CPA2) were examined for their role in building a Parkinson's disease diagnostic model. The discrimination ability of a single protein for the disease group and the control group was studied (as shown in the discrimination AUC in Table 7), and 2 or 3 or 4 of the 5 proteins were selected to train a diagnostic model, and the discrimination ability was calculated in the test set (as shown in AUC in Table 8), and as shown in Figure 2 ​The model trained with the five proteins has a discrimination efficiency of 91.7% for Parkinson's disease. It should be noted that the above is to use logistic regression to construct the model and use the Bayesian information criterion method to analyze the model, and then draw the ROC curve (receiver operating characteristic curve) according to the obtained protein markers, and calculate the area under the curve (AUC) as the discrimination model efficiency evaluation parameter. As can be seen from the results in Tables 7 and 8, each single protein (IL10, CX3CL1, DPEP1, CLEC10A, CPA2) has a certain ability to distinguish Parkinson's disease, and the diagnostic models constructed with ① CPA2, DPEP1 or ② CPA2, DPEP1, CX3CL1 or ③ CPA2, DPEP1, CX3CL1, CLEC10A or ④ CPA2, DPEP1, CX3CL1, CLEC10A, IL10 have a model accuracy (i.e. AUC) of 0.85 or more. Therefore, IL10, CX3CL1, DPEP1, CLEC10A, and CPA2 are indeed protein markers with strong correlation with Parkinson's disease diagnosis, and any one or more of the five proteins can be used to establish a Parkinson's disease diagnosis model, and an ideal effect can be obtained.

[0114] In particular, the accuracy (AUC) of the protein combination formed by DEEP1 and CPA2 for predicting Parkinson's disease is also 0.85 (see Table 8), which is sufficient to support DEEP1 and / or CPA2 as a risk marker for predicting Parkinson's disease with high sensitivity and high specificity. DEEP1 and / or CPA2, optionally combined with any one or more of the other 16 proteins, can be used to construct a diagnostic model to better predict Parkinson's disease.

[0115] Table 7 Optimal set of protein markers for disease group and control group samples

[0116]

[0117] Table 8 Combination of the top 5 most important protein markers

[0118]

[0119]

[0120] The optimal set of CPA2, DPEP1, CX3CL1, CLEC10A, and IL10 performs best, with an area under the ROC curve (AUC) of 91.7% Figure 2 .

[0121] From the above description, it can be seen that the above-mentioned embodiments of the present application achieve the following technical effects: the present application uses one or more of specific protein markers to prepare a product for diagnosing or predicting Parkinson's disease, which can diagnose or predict the condition or risk of Parkinson's disease in a subject.

[0122] The above only describes the preferred embodiments of the present application and is not intended to limit the present application. Those skilled in the art can make various modifications and changes to the present application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.

Claims

1. Use of a protein marker for the manufacture of a product for the diagnosis of Parkinson or the prediction of Parkinson's disease, characterized in that, The protein biomarkers are selected from any one of the following groups: (1) CLEC10A, CPA2, NTRK3 and SNCG; or (2) CLEC10A, CPA2, NTRK3 and DPEP1; or (3) CLEC10A, CPA2, NTRK3, DPEP1 and SNCG; or (4) IL10, CLEC10A, CPA2, NTRK3 and SNCG; or (5) CLEC10A, NTRK3 and DPEP1; or (6) CX3CL1, CXCL6, IL10, CLEC10A and DPEP1; or (7) CX3CL1, CXCL6, IL.17C, CLEC10A and DPEP1; or (8) CX3CL1, CXCL6, CLEC10A and DPEP1; or (9) CX3CL1, CXCL6, CLEC10A, DPEP1 and PSME1; or (10) CX3CL1, CXCL6, FGF.23, CLEC10A and DPEP1; or (11) IL10, CLEC10A, CPA2 and DPEP1; or (12) IL10, CLEC10A and DPEP1; or (13) CPA2, DPEP1, CX3CL1 and CLEC10A; or (14) CPA2, DPEP1, CX3CL1, CLEC10A and IL10.

2. Use according to claim 1, characterized in that, The application includes using detection reagents to detect the content of the protein marker.

3. Use according to claim 2, characterized in that, The methods for detecting the content of the protein markers include Olink technology, Western blotting, or enzyme-linked immunosorbent assay.

4. Use according to claim 2, characterized in that, The detection reagent is an antibody against the protein marker.

5. Use according to claim 4, characterized in that, The antibody is set on a solid support.

6. Use according to claim 5, characterized in that, The solid-phase carrier includes enzyme-labeled plates, membrane carriers, or microspheres.

7. Use according to claim 6, characterized in that, The membrane carrier is selected from nitrocellulose membrane, glass cellulose membrane or nylon membrane.

8. Use according to claim 4, characterized in that, The antibody is a monoclonal antibody or a polyclonal antibody.

9. The use according to any one of claims 1 to 8, characterized in that, Related products for this application include reagent kits or electronic prediction devices.

10. A kit for diagnosing or predicting Parkinson's disease in a subject, characterized in that, The kit includes a detection reagent for detecting the content of a protein biomarker, wherein the protein biomarker is selected from any one of the following groups: (1) CLEC10A, CPA2, NTRK3 and DPEP1; or (2) CLEC10A, CPA2, NTRK3, DPEP1 and SNCG; or (3) CLEC10A, NTRK3 and DPEP1; or (4) CX3CL1, CXCL6, IL10, CLEC10A and DPEP1; or (5) CX3CL1, CXCL6, IL.17C, CLEC10A and DPEP1; or (6) CX3CL1, CXCL6, CLEC10A and DPEP1; or (7) CX3CL1, CXCL6, CLEC10A, DPEP1 and PSME1; or (8) CX3CL1, CXCL6, FGF.23, CLEC10A and DPEP1; or (9) IL10, CLEC10A, CPA2 and DPEP1; or (10) IL10, CLEC10A and DPEP1; or (11) CPA2, DPEP1, CX3CL1 and CLEC10A; or (12) CPA2, DPEP1, CX3CL1, CLEC10A and IL10. 11.An electronic device for diagnosing or predicting Parkinson’s disease, comprising: The electronic device comprises: an acquisition detection module configured to acquire a biological sample, detect the biological sample, and obtain the content of protein markers in the biological sample; a diagnosis or prediction module configured to establish a regression equation using a logistic regression model, a support vector machine model or a naive Bayes model and the content of the protein markers, construct a model, and then output a prediction result according to the model; The protein markers are selected from any one of the following groups: (1) CLEC10A, CPA2, NTRK3 and DPEP1; or (2) CLEC10A, CPA2, NTRK3, DPEP1 and SNCG; or (3) CLEC10A, NTRK3 and DPEP1; or (4) CX3CL1, CXCL6, IL10, CLEC10A and DPEP1; or (5) CX3CL1, CXCL6, IL.17C, CLEC10A and DPEP1; or (6) CX3CL1, CXCL6, CLEC10A and DPEP1; or (7) CX3CL1, CXCL6, CLEC10A, DPEP1 and PSME1; or (8) CX3CL1, CXCL6, FGF.23, CLEC10A and DPEP1; or (9) IL10, CLEC10A, CPA2 and DPEP1; or (10) IL10, CLEC10A and DPEP1; or (11) CPA2, DPEP1, CX3CL1 and CLEC10A; or (12) CPA2, DPEP1, CX3CL1, CLEC10A and IL10. The biological sample is a plasma sample. 12.The electronic device of claim 11, wherein, The prediction result output according to the model is a probability value output by using the predict function of type='response', and the probability value≥0.5 indicates that the subject has the Parkinson's disease or has a high risk, and the probability value<0.5 indicates that the subject does not have the Parkinson's disease or has a low risk.

Images