Parkinson's disease specific plasma metabolism marker and application thereof

Abstract

The invention belongs to the technical field of biomedicine, and discloses a Parkinson's disease specific plasma metabolism marker and application thereof.The plasma metabolism marker comprises uracil, kynurenine, phenolsulfonic acid, p-cresol, di-high gamma-linolenic acid, atorvastatin, 2-methoxybenzoic acid, carabicycline, nitrogen hydroxyl 2-amino-3, 5-di-tert-butyl-3, 5-di-tert-butyl-3, 5-di-tert-butyl-3, 5-di-tert-butyl-3, 5-di-tert-butyl-3, 5- The invention relates to an anti-inflammatory agent which is prepared from 2, 8-dimethylimidazo quinoxaline, dihydroxy phenethyl alcohol O-sulfate, racemization-4-hydroxy-4-O-(beta-D-glucuronate)-all-trans retinol acetate, fructosyl valine, thiaxanthene and acetylpyridine methanol. It is proved for the first time that the 14 plasma metabolites are associated with the Parkinson's disease, a Parkinson's disease prediction method is constructed based on the markers, the prediction result is reliable, and a new way is provided for diagnosis of the Parkinson's disease.

Description

Technical Field

[0001] The present invention belongs to the field of biomedical technology, and specifically relates to Parkinson's disease specific plasma metabolic markers and their application in preparing Parkinson's disease diagnostic reagents and constructing Parkinson's disease prediction systems. Background Art

[0002] Parkinson's disease is the second largest neurodegenerative disease in the world. It is more common in the elderly population, with a prevalence of about 2-3% in people over 65 years old. The main pathological features of Parkinson's disease are the death of dopaminergic neurons in the substantia nigra and the appearance of Lewy bodies. The main clinical features of Parkinson's disease are divided into motor symptoms and non-motor symptoms. Motor symptoms generally manifest as resting tremor, muscle rigidity and bradykinesia; non-motor symptoms mainly manifest as abnormal smell, gastrointestinal disorders, cognitive impairment, etc. The specific causes and pathogenic mechanisms of Parkinson's disease are complex, but it is generally believed that it is related to genetics and the environment. At present, there is no target or hypothesis that can well explain the cause of Parkinson's disease and predict the occurrence of Parkinson's disease.

[0003] At present, the clinical diagnosis of Parkinson's disease mainly relies on the clinical evaluation of the patient's motor symptoms, such as resting tremor and bradykinesia. This makes the early diagnosis of Parkinson's disease more subjective and difficult. In addition, the clinical symptom characteristics of Parkinson's disease are similar to those of other diseases such as Parkinson's disease syndrome and essential tremor, which leads to a longer diagnosis period for Parkinson's disease and is more prone to misdiagnosis. At the same time, compared with early patients, the drug treatment effect of patients in the middle and late stages is also poor. Therefore, there is an urgent need for a new diagnostic method for Parkinson's disease, which can serve as an indicator of the onset of early Parkinson's disease and assist doctors in diagnosis.

[0004] Existing evidence shows that Parkinson's disease is a systemic, multi-transmitter disease, and peripheral tissues and systems may play a very important role in the onset of the disease. Previous studies have shown that compared with healthy people, other immune cells such as T lymphocytes, B lymphocytes, cytokines, and natural killer cells in the peripheral blood are dysregulated in Parkinson's disease groups. These results indicate that factors such as the level of inflammation in peripheral blood play a very important role in the pathogenesis of Parkinson's disease, and that peripheral blood may contain potential Parkinson's disease biomarkers and therapeutic targets. Metabolites in the human body are closely related to genes and proteins. The composition of metabolites in a cell or tissue reflects the changes it is sending, and is a direct substance that reflects the environment in which the organism is located and the health status. With the advancement of omics technology, it has been possible to quantify all molecules of the body at one time, allowing researchers to explore the pathological and molecular characteristics of the disease from a systemic level. Metabolomics, as a technical means to comprehensively study metabolites in organisms, has been widely used in early diagnosis and treatment monitoring of diseases. Through quantitative and qualitative analysis of metabolites in organisms, metabolomics technology can discover disease-specific metabolic markers, providing new ideas and methods for early diagnosis of diseases. Summary of the invention

[0005] Traditional Parkinson's disease diagnostic methods have problems such as inaccurate diagnosis, cumbersome operation, and high cost, which seriously restrict the early detection and treatment of Parkinson's disease. In response to this problem, the present invention aims to combine metabolomics technology and use non-targeted metabolomics technology to find Parkinson's disease-specific metabolic markers, providing new ideas for the diagnosis and treatment of Parkinson's disease.

[0006] In order to achieve the above object, the technical solution of the present invention is as follows:

[0007] The first aspect of the present invention provides the use of plasma metabolite markers and expression detection reagents thereof in the preparation of a Parkinson's disease diagnostic kit, which comprises reagents for detecting plasma metabolite markers in the plasma of a subject, and the plasma metabolite markers include phenolsulfonic acid, atorvastatin, 2-methoxybenzoic acid and dihydroxyphenylethanol O-sulfate.

[0008] Preferably, in the above application, the plasma metabolic markers may also include at least one or more of the following: uracil, kynurenine, p-cresol, dihomo-gamma linolenic acid, carbacycline, nitrogen hydroxy 2-amino-3,8-dimethylimidazoquinoxaline, racemic-4-hydroxy-4-O-(β-D-glucuronic acid ester)-all-trans retinol acetate, fructosyl valine, thioxanthene and acetylpyridine methanol.

[0009] The present invention discovers for the first time that the above 14 plasma metabolites are imbalanced in Parkinson's disease patients, and compared with healthy people, the contents of 9 metabolites, including p-cresol, dihomo-gamma-linolenic acid, atorvastatin, 2-methoxybenzoic acid, carbacycline, nitrogen hydroxyquinoxaline, dihydroxyphenyl glycol O-sulfate, fructosyl valine, and acetylpyridine methanol, in plasma are significantly increased in Parkinson's disease patients, while the contents of 5 metabolites, including uracil, kynurenine, phenolsulfonic acid, racemic-4-hydroxy-4-O-(β-D-glucuronic acid ester)-all-trans retinol acetate, and thioxanthene, are significantly reduced in Parkinson's disease patients. Furthermore, the random forest disease prediction model showed that using the above 14 metabolites as markers, the model can well distinguish between Parkinson's disease patients and healthy people, and different combinations of these 14 metabolites also have certain capabilities and predictive accuracy for the diagnosis of Parkinson's disease. Among them, the combination of phenolsulfonic acid, atorvastatin, 2-methoxybenzoic acid and dihydroxyphenylethanol O-sulfate has the highest predictive performance for Parkinson's disease.

[0010] The second aspect of the present invention provides the use of plasma metabolic markers in screening drugs for treating Parkinson's disease, wherein the plasma metabolic markers include uracil, kynurenine, phenolsulfonic acid, p-cresol, dihomo-gamma linolenic acid, atorvastatin, 2-methoxybenzoic acid, carbacycline, nitrogen hydroxy 2-amino-3,8-dimethylimidazoquinoxaline, dihydroxyphenylethanol O-sulfate, racemic-4-hydroxy-4-O-(β-D-glucuronic acid ester)-all-trans retinol acetate, fructosyl valine, thioxanthene and acetylpyridine methanol.

[0011] The third aspect of the present invention provides the use of plasma metabolic markers in constructing a Parkinson's disease prediction system, wherein the plasma metabolic markers include phenolsulfonic acid, atorvastatin, 2-methoxybenzoic acid and dihydroxyphenylethanol O-sulfate; preferably, the plasma metabolic markers may also include any one or more of the following: uracil, kynurenine, p-cresol, dihomo-gamma linolenic acid, carbacycline, nitrogen hydroxy 2-amino-3,8-dimethylimidazoquinoxaline, racemic-4-hydroxy-4-O-(β-D-glucuronic acid ester)-all-trans retinol acetate, fructosyl valine, thioxanthene and acetylpyridine methanol.

[0012] Preferably, in the above application, the expression of plasma metabolic markers is used as a characteristic variable, and a logistic regression machine learning model is used to train the model using a five-fold cross validation method. For example, in one embodiment of the present invention, the Logistic Regression function in sklearn is specifically used for modeling, wherein the parameters random_state = 1, solver = "saga", and penalty = "l2". Experimental data show that when these 14 metabolites are used as markers to predict the onset of Parkinson's disease, they not only have high AUC values, specificity and sensitivity in the training set, but also have high AUC values, specificity and sensitivity in the validation set, indicating that the prediction results are reliable and have application value in diagnosing Parkinson's disease.

[0013] Preferably, in the above application, the expression level of plasma metabolic markers is obtained by high performance liquid chromatography tandem mass spectrometry.

[0014] The fourth aspect of the present invention provides a Parkinson's disease prediction system, which uses the expression level of plasma metabolite markers in the plasma of the subject as an input variable, and the plasma metabolite markers include uracil, kynurenine, phenolsulfonic acid, p-cresol, dihomo-gamma linolenic acid, atorvastatin, 2-methoxybenzoic acid, carbacycline, nitrogen hydroxy 2-amino-3,8-dimethylimidazoquinoxaline, dihydroxyphenylethanol O-sulfate, racemic-4-hydroxy-4-O-(β-D-glucuronic acid ester)-all-trans retinol acetate, fructosyl valine, thioxanthene and acetylpyridine methanol.

[0015] Compared with the prior art, the beneficial effects of the present invention are:

[0016] The present invention is the first to use non-targeted metabolomics technology to search for metabolic markers of Parkinson's disease. Although this method cannot completely detect all metabolites in plasma, it is currently the metabolite detection method with the highest coverage and can quantify metabolites in tissues to the greatest extent. Therefore, the metabolic markers provided by the present invention are highly accurate, and the 14 metabolites provided by the present invention have not been reported to be associated with Parkinson's disease in previous studies.

[0017] Compared with other existing Parkinson's disease diagnostic schemes, the plasma metabolic marker combination and machine learning method provided by the present invention have a higher AUC value and higher sensitivity for the prediction of Parkinson's disease, that is, the prediction results are more reliable.

[0018] The Parkinson's disease prediction system developed by the present invention based on the quantitative detection of plasma metabolites is less invasive, simple and accurate. At the same time, the plasma metabolites disclosed in the present invention can be used as targets for the development of drugs to alleviate Parkinson's disease. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 Flow chart for developing and validating metabolite molecules that can serve as Parkinson's disease markers for the present invention;

[0020] Figure 2 This is a diagram showing the difference analysis results of the metabolite molecules uracil, kynurenine, phenolsulfonic acid, p-cresol, dihomo-γ-linolenic acid, atorvastatin, 2-methoxybenzoic acid and carbacycline in Example 1;

[0021] Figure 3 This is a graph showing the difference analysis results of the metabolite molecules nitrogen hydroxy 2-amino-3,8-dimethylimidazoquinoxaline, dihydroxyphenylethanol O-sulfate, racemic-4-hydroxy-4-O-(β-D-glucuronic acid ester)-all-trans-retinol acetate, fructosyl valine, thioxanthene, and acetylpyridinemethanol in Example 1;

[0022] Figure 4 VIP value statistics of 14 differential metabolite molecules in Example 1;

[0023] Figure 5 This is the ROC curve in Example 3, where the abscissa is 1-specificity and the ordinate is sensitivity. DETAILED DESCRIPTION

[0024] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which the present invention belongs. The term "comprises" and any variations thereof in the specification and claims of the present invention are intended to cover non-exclusive inclusions.

[0025] Since the diagnosis of Parkinson's disease is still mainly based on clinical symptoms, which seriously restricts the early detection and treatment of Parkinson's disease, the present invention hopes to find new possibilities for the diagnosis of Parkinson's disease by introducing metabolomics technology. Figure 1 As shown, the scheme of the present invention comprises the following steps:

[0026] (1) Blood samples from healthy people and Parkinson's disease patients were obtained respectively, and metabolites in plasma were analyzed by high performance liquid chromatography tandem-mass spectrometry (LC-MS / LC); the detection data were processed, and the data processing method included but was not limited to preprocessing, normalization, quality control, etc., so as to obtain the expression level of each compound in each sample.

[0027] (2) Based on the expression results, the generalized linear mixed model was used for differential analysis to screen out the differentially metabolized molecules that met the criteria. Then, orthogonal partial least squares discriminant analysis (OPLS-DA) was used to screen out the molecules with VIP values ​​(Variable Importance in Projection) that met the requirements as markers.

[0028] (3) Based on the above-mentioned metabolite molecules used as markers, a logistic regression algorithm was used with a five-fold cross-validation method to perform data training and internal validation in the test set.

[0029] (4) In addition, blood samples from healthy people and Parkinson's disease patients were collected to obtain the expression levels of the above metabolite molecules, which were substituted into the above model for verification, and the AUC value and ROC curve were calculated to evaluate the stability of the model performance.

[0030] The present invention performs non-targeted quantification of plasma metabolites in positive ion mode and negative ion mode respectively, and uses difference analysis and orthogonal partial least squares discriminant analysis in turn to screen metabolic markers that can reflect the disease state by comparing the differences between groups. The present invention first discovered that 14 plasma metabolites are dysregulated in Parkinson's disease patients, specifically: uracil, kynurenine, phenolsulfonic acid, p-cresol, dihomo-γ-linolenic acid, atorvastatin, 2-methoxybenzoic acid, carbacycline, nitrogen hydroxy 2-amino-3,8-dimethylimidazoquinoxaline, dihydroxyphenyl glycol O-sulfuric acid, racemic 4-hydroxy 4Oβ-D-glucuronide all-trans retinyl acetate, fructosyl valine, thioxanthene and acetylpyridine methanol. Using these 14 metabolites as characteristic variables, a logistic regression machine learning model was used to predict the disease. In the training set, five-fold cross validation was used to find that these 14 metabolites had high AUC values, specificity, and sensitivity for predicting the onset of Parkinson's disease. At the same time, they were verified in another validation set and found that they also had high AUC values, specificity, and sensitivity in the validation set.

[0031] The technical solution of the present invention will be described clearly and completely below in conjunction with the embodiments and drawings. It should be understood that the specific implementation methods described herein are only used to illustrate and explain the present invention and are not used to limit the present invention.

[0032] In the following examples, unless otherwise specified, all methods are conventional methods; the reagents and materials described, unless otherwise specified, can be obtained from commercial sources.

[0033] Example 1

[0034] In this case, a plasma target metabolite combination that can be used to develop a prediction system or detection reagent for Parkinson's disease was obtained through non-targeted metabolomics technology, which specifically includes the following processes:

[0035] (1) Sample collection.

[0036] In the Parkinson's disease group, blood samples were collected from 25 Parkinson's disease patients who visited the Department of Neurology of the First People's Hospital of Xiangyang City, Hubei Province. The inclusion criteria were: ① patients met the 2015 International Movement Disorder Society diagnostic criteria for primary Parkinson's disease; ② patients had no secondary Parkinson's syndrome, Parkinson's plus syndrome, and other neurodegenerative diseases; ③ patients had no malignant tumors, gastrointestinal diseases, liver and kidney dysfunction, gout, and other serious diseases.

[0037] For the healthy group, blood samples were collected from 25 healthy spouses, and the samples also came from the First People's Hospital of Xiangyang City, Hubei Province. The inclusion criteria were: ① Living with Parkinson's disease patients for more than 30 years; ② Being in good health, without any neurological diseases or other major diseases.

[0038] (2) Plasma extraction and mass spectrometry detection.

[0039] All participants had 10 mL of venous blood drawn on an empty stomach on the day of consultation or in the early morning of the next day. The blood was placed in an EDTA anticoagulant tube and centrifuged at 3000×g (4°C, 10 minutes) to obtain plasma. The blood was quickly placed in liquid nitrogen at -196°C and, after being completely frozen, was transferred to a -80°C refrigerator for storage.

[0040] Metabolite extraction: Slowly thaw the above samples at 4°C, take 100 μL, add 300 μL extraction solution (methanol: acetonitrile = 2:1, -20°C) and 10 μL internal standard, mix for 1 minute, let stand at -20°C for 2 hours, then centrifuge at 4000 rcf at 4°C for 20 minutes; take 300 μL supernatant, place it in a freeze vacuum concentrator to dry, add 150 μL reconstitution solution (methanol: H2O = 1:1) for reconstitution, and take the supernatant and place it in the loading bottle. Take 10 μL of the supernatant of each sample and mix it into a quality control (QC) sample to evaluate the repeatability and stability of the chromatography-mass spectrometry analysis process.

[0041] Mass spectrometry detection: In this experiment, Waters 2DUPLC (Waters, USA) was used in conjunction with Q Exactive high-resolution mass spectrometer (Thermo Fisher Scientific, USA) for separation and detection of metabolites. The chromatographic conditions and mass spectrometry conditions are as follows:

[0042] The chromatographic column used was a BEH C18 column (1.7 μm, 2.1×100 mm, Waters, USA). The mobile phases in the positive ion mode were an aqueous solution containing 0.1% formic acid (liquid A) and 100% methanol containing 0.1% formic acid (liquid B), and the mobile phases in the negative ion mode were an aqueous solution containing 10 mM ammonium formate (liquid A) and 95% methanol containing 10 mM ammonium formate (liquid B). The following gradient was used for elution: 0-1 min, 2% B solution; 1-9 min, 2-98% B solution; 9-12 min, 98% B solution; 12-12.1 min, 98% B solution-2% B solution; 12.1-15 min, 2% B solution. The flow rate was 0.35 mL / min, the column temperature was 45°C, and the injection volume was 5 μL.

[0043] The Q Exactive mass spectrometer was used for primary and secondary mass spectrometry data acquisition. The mass-to-nuclear ratio range of the mass spectrometry scan was 70-1050, the primary resolution was 70,000, and the AGC was 3e 6 The maximum injection time (IT) was 100 milliseconds. According to the parent ion intensity, Top3 was selected for fragmentation and secondary information was collected. The secondary resolution was 17,500 and the AGC was 1e 5 , the maximum injection time is 50 milliseconds, and the fragmentation energy (stepped nce) is set to: 20, 40, 60eV. The ion source (ESI) parameter settings are: sheath gas flow rate is 40, auxiliary gas flow rate is 10, spray voltage (|KV|) is 3.80 in positive ion mode and 3.20 in negative ion mode, ion transfer tube temperature (Capillary temp) is 320℃, and auxiliary gas heater temperature (Aux gas heater temp) is 350℃.

[0044] (3) Metabolomics data processing.

[0045] The original data (raw file) was input into Compound Discoverer 3.1 (Thermo Fisher Scientific, USA) for data processing, which mainly included peak extraction, intra-group and inter-group retention time correction, adduct ion merging, missing value filling, background peak marking, etc. Finally, metabolite annotation was performed using multiple databases such as mzCloud, ChemSpider, HMDB, and KEGG.

[0046] The above results were imported into metaX for the next step of data processing, which mainly included: ① normalization, using the PQN (Probabilistic Quotient Normalization) method to normalize the data and obtain the relative peak area value; ② batch effect correction, using QC-RLSC (Quality control-based robust LOESS signal correction) to correct the batch effect; ③ all QC samples in which the coefficient of variation of the relative peak area (Coefficient of Variation) was greater than 30% were eliminated.

[0047] After the above steps, we obtained the expression level of each compound in each sample.

[0048] (4) Metabolome data analysis.

[0049] Data preprocessing: log2 values ​​of each metabolite detected by mass spectrometry were taken.

[0050] Identification of differential metabolites: We used a generalized linear mixed model with group (Parkinson's disease group and healthy group) as fixed effect variable, sex, age, and levodopa equivalent dose as random effect variables, and used the glmer function in the R package lmer to calculate the generalized linear mixed model (glmer(Group~metabolite+(1|Age)+(1|Sex)+(1|LEDD)). Finally, the BH (Benjamini & Hochberg) method was used for multiple testing correction. Finally, the differential metabolite molecules were screened out with p.adjusted<0.05 as the condition.

[0051] Feature selection: Orthogonal partial least squares discriminant analysis was used to screen out the molecules with VIP values ​​greater than 1 among the differentially metabolized molecules obtained above as markers.

[0052] Finally, we obtained 14 metabolite molecules, as shown in Table 1:

[0053] Table 1 Metabolite molecular information

[0054]

[0055] Note: / indicates that there is no corresponding ID in the corresponding database.

[0056] Figure 2-3 and Figure 4The following are the difference analysis results and VIP value results of these 14 metabolite molecules, among which Uridine is uracil, L-Kynurenine is kynurenine, p-Phenolsulfonic acid is phenolsulfonic acid, 4-Methylphenol is p-cresol, Dihomo-gamma-linolenate is dihomo-gamma-linolenate, Atorvastatin is atorvastatin, 2-Methoxybenzoic acid is 2-methoxybenzoic acid, Carbacyclin is carbacyclin, N-Hydroxy-MeIQx is nitrogen hydroxy 2-amino-3,8-dimethylimidazoquinoxaline, 3,4-Dihydroxyphenylglycol O-sulfate is dihydroxyphenylethanol O-sulfate, rac-4-Hydroxy-4-O-(beta-D-glucuronide)-all-trans-retinyl acetate is racemic-4-hydroxy-4-O-(β-D-glucuronide)-all-trans-retinyl acetate, Fructosyl Valine is fructosyl valine, Thioxanthene is thioxanthene, and 1H-indol-3-yl(pyridin-2-yl)methanol is acetylpyridinium methanol.

[0057] At the same time, from the above results, it can be seen that compared with healthy people, the levels of 9 metabolites including p-cresol, dihomo-gamma-linolenic acid, atorvastatin, 2-methoxybenzoic acid, carbacycline, nitrogen hydroxyquinoxaline, dihydroxyphenylglycol O-sulfate, fructosyl valine, and acetylpyridine methanol in plasma were significantly increased in people with Parkinson's disease, while the levels of 5 metabolites including uracil, kynurenine, phenolsulfonic acid, racemic-4-hydroxy-4-O-(β-D-glucuronic acid ester)-all-trans-retinol acetate, and thioxanthene were significantly decreased in people with Parkinson's disease.

[0058] Example 2

[0059] Based on the 14 metabolite molecules that can be used as markers provided in Example 1, this example provides a method for constructing a disease prediction system, which is as follows:

[0060] This example takes into account all the permutations and combinations of metabolite molecules, uses a logistic regression algorithm, and adopts a five-fold cross validation method to train and internally validate data in a test set (i.e., the detection data of Example 1). This example specifically uses the Logistic Regression function in sklearn to model, with parameters random_state=1, solver="saga", penalty="l2", and uses a five-fold cross validation (Repeated Stratified KFold(n_splits=5,n_repeats=100,random_state=1)) method. Two indicators are used to evaluate the effect of the model at the same time, one is sensitivity, and the other is the area under the ROC (Receiver Operating Characteristic Curve, receiver operating characteristic) curve AUC (Area under the Curve) value. Sensitivity and specificity are mainly calculated by the true positive number (TP), false positive number (FP), true negative number (TN), and false negative number (FN) in the results, where sensitivity=TP / (TP+FN) and specificity=TN / (TN+FP).

[0061] Among them, the detection results of some combinations are shown in Table 2:

[0062] Table 2 Test set evaluation results

[0063]

[0064]

[0065]

[0066] Example 3

[0067] In this example, samples from 60 patients and their healthy spouses were collected (i.e., blood samples from 30 Parkinson's disease patients and 30 spouses, with the same sources and inclusion criteria as in Example 1), and the expression levels of 14 metabolite molecules in each sample were obtained with reference to Example 1.

[0068] Then, the model was trained using the logistic regression algorithm with all samples in the test set (same as in Example 2), and finally used to predict the situation in the validation set, and its AUC value and sensitivity were calculated to evaluate the stability of the model performance. The results are shown in Table 3 (Table 3 only lists the top 100 metabolite combinations with the highest AUC values ​​obtained in this example):

[0069] Table 3. Evaluation results of validation set

[0070]

[0071]

[0072]

[0073] In summary, the present invention has proved for the first time that the 14 metabolite molecules in Table 1 are associated with Parkinson's disease, and the random forest disease prediction model shows that with these 14 metabolites as markers, the model can well distinguish Parkinson's disease patients from healthy people. In the test set, the average AUC value of the model was 0.97, the sensitivity was 0.86, and the specificity was 0.97; in the external validation set, the AUC value of the model reached 0.92 and the sensitivity reached 0.96. And the different combinations of these 14 metabolites also have certain capabilities and prediction accuracy for the diagnosis of Parkinson's disease, among which the combination of M3+M7+M10+M6 has the highest prediction performance for Parkinson's disease, and its AUC value in the validation set reached 0.95. Compared with other patents related to Parkinson's disease diagnosis (such as CN113960199A, CN 111024843 A, CN 114085899 A, CN 117476111 A), the marker combination and machine learning method of the present invention have a higher AUC value and higher sensitivity for the prediction of Parkinson's disease, and the prediction results for Parkinson's disease are more reliable.

[0074] It should be noted that the above embodiments are only part of the embodiments of the present invention rather than all the embodiments, and are only used to illustrate the technical solution of the present invention rather than to limit it. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

Claims

1. The use of plasma metabolic markers and detection reagents thereof in the preparation of a Parkinson's disease diagnostic kit, characterized in that: The plasma metabolite markers include phenolsulfonic acid, atorvastatin, 2-methoxybenzoic acid and dihydroxyphenylethanol O-sulfate.

2. The use according to claim 1, characterized in that: The plasma metabolic markers also include any one or more of the following: uracil, kynurenine, p-cresol, dihomo-gamma linolenic acid, carbacycline, nitrogen hydroxy 2-amino-3,8-dimethylimidazoquinoxaline, racemic-4-hydroxy-4-O-(β-D-glucuronic acid ester)-all-trans retinol acetate, fructosyl valine, thioxanthene and acetylpyridine methanol.

3. The use according to claim 1, characterized in that: The expression levels of the plasma metabolic markers in the plasma of the subjects were detected, and the expression levels of p-cresol, dihomo-gamma linolenic acid, atorvastatin, 2-methoxybenzoic acid, carbacycline, nitrogen hydroxyquinoxaline, dihydroxyphenyl glycol O-sulfuric acid, fructosyl valine, and acetylpyridine methanol in the plasma of Parkinson's disease patients were higher than those in healthy people, while the expression levels of uracil, kynurenine, phenolsulfonic acid, racemic-4-hydroxy-4-O-(β-D-glucuronic acid ester)-all-trans retinyl acetate, and thioxanthene were lower than those in healthy people.

4. The use of plasma metabolic markers in screening drugs for treating Parkinson's disease, characterized in that: The invention comprises reagents for detecting plasma metabolic markers in the plasma of a subject, wherein the plasma metabolic markers include uracil, kynurenine, phenolsulfonic acid, p-cresol, dihomo-gamma linolenic acid, atorvastatin, 2-methoxybenzoic acid, carbacycline, nitrogen hydroxy 2-amino-3,8-dimethylimidazoquinoxaline, dihydroxyphenylethanol O-sulfate, racemic-4-hydroxy-4-O-(β-D-glucuronic acid ester)-all-trans retinol acetate, fructosyl valine, thioxanthene and acetylpyridine methanol.

5. Application of plasma metabolic markers in constructing a Parkinson's disease prediction system, characterized in that: The plasma metabolite markers include phenolsulfonic acid, atorvastatin, 2-methoxybenzoic acid and dihydroxyphenylethanol O-sulfate.

6. The use according to claim 5, characterized in that: Plasma metabolic markers also include at least one or more of the following: uracil, kynurenine, p-cresol, dihomo-γ-linolenic acid, carbacycline, nitrogen hydroxy 2-amino-3,8-dimethylimidazoquinoxaline, racemic-4-hydroxy-4-O-(β-D-glucuronic acid ester)-all-trans retinyl acetate, fructosyl valine, thioxanthene and acetylpyridine methanol.

7. The use according to claim 5 or 6, characterized in that: The expression levels of plasma metabolic markers were used as characteristic variables, and a logistic regression machine learning model was used to train the model using the five-fold cross-validation method.

8. The use according to claim 7, characterized in that: The expression levels of the plasma metabolic markers are obtained by high performance liquid chromatography tandem mass spectrometry.

9. The use according to claim 7, characterized in that: The Logistic Regression function in sklearn is used for modeling, with parameters random_state = 1, solver = "saga", and penalty = "l2".

10. A Parkinson's disease prediction system, characterized in that: The expression levels of plasma metabolite markers in the plasma of the subjects are used as input variables, and the plasma metabolite markers include uracil, kynurenine, phenolsulfonic acid, p-cresol, dihomo-γ-linolenic acid, atorvastatin, 2-methoxybenzoic acid, carbacycline, nitrogen hydroxy 2-amino-3,8-dimethylimidazoquinoxaline, dihydroxyphenylethanol O-sulfate, racemic-4-hydroxy-4-O-(β-D-glucuronic acid ester)-all-trans retinol acetate, fructosyl valine, thioxanthene and acetylpyridine methanol.

Images