Urine biomarker for yin deficiency syndrome of Alzheimer's disease and application of urine biomarker
By combining non-targeted metabolomics and machine learning algorithms, urine biomarkers for yin deficiency syndrome of Alzheimer's disease were determined, and diagnostic models were constructed, which solved the problem of irregular diagnostic methods in the existing technology, and achieved the goal of Alzheimer's disease. Early and efficient diagnosis of Yin deficiency syndrome of Zheimer's disease, with high sensitivity and specificity, and is suitable for urine sample detection.
Patent Information
- Application Number
- CN202411280757.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-13
- Publication Date
- 2025-01-10
AI Technical Summary
The existing technology lacks effective diagnostic methods to distinguish between yin deficiency syndrome and non-yin deficiency syndrome in Alzheimer's disease, resulting in limited clinical and scientific research work in traditional Chinese medicine, and has not formed diagnostic technical specifications, with many subjective factors, few objective indicators, and dialectical results. Issues such as large individual differences.
Non-targeted metabolomics technology combined with machine learning algorithms to determine urine biomarkers for yin deficiency syndrome of Alzheimer's disease, and detect hesperin, kaempferol, adrenaline, uridine monophosphate and amino groups through urine samples. The combination of butyric acid builds a diagnostic model, provides early auxiliary diagnostic kits and diagnostic models.
It realizes early auxiliary diagnosis of yin deficiency syndrome of Alzheimer's disease, has strong operability, simple model construction, high sensitivity and excellent specificity, and is suitable for urine detection without additional tissue samples collection, reducing the risk of trauma and having high Clinical application value.
Smart Images
Figure CN119291060A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of clinical test and diagnosis applications, and in particular to a urine biomarker for Yin deficiency syndrome of Alzheimer's disease and an application thereof. Background Art
[0002] Alzheimer's disease (AD) is the most common type of dementia. It has an insidious onset and is difficult to diagnose in the early stages, leading to cognitive impairment, mental and behavioral problems, and loss of social and living functions, placing a heavy economic and social burden on the public health system.
[0003] Yin deficiency syndrome refers to the deficiency of yin fluid in the human body, which leads to the decline of its functions of moistening and nourishing, or the inability of yin to control yang, resulting in excessive yang, and the main manifestations of deficiency heat syndrome are dry mouth and throat, hot flashes and night sweats. According to traditional Chinese medicine, Alzheimer's disease belongs to "dementia disease", which is mostly caused by deficiency of heart, liver, spleen and kidney in old age, inability to transport and transform essence, endogenous phlegm, blood stasis, fire and poison, marrow reduction and brain loss, lack of nourishment for the spirit, and dysfunction of the spirit, resulting in dementia. The collation of ancient Chinese medicine literature on dementia found that deficiency syndrome is the main pathogenesis of dementia, especially "yin deficiency". The statistics of constitution and syndrome type of AD patients also found that yin deficiency syndrome is the main syndrome type of AD.
[0004] At present, there are few studies on the evaluation methods of Alzheimer's Disease Yin Deficiency Syndrome (ADY) and Alzheimer's Disease Not-Yin Deficiency Syndrome (ADNY), and the technical specifications for ADY diagnosis have not yet been formed. The existing evaluation methods have many subjective factors, few objective indicators, and large individual differences in syndrome differentiation results, which affect the development of TCM clinical and scientific research. Based on urine metabolomics, searching and discovering urine biomarkers of ADY can provide a basis for revealing the core pathogenesis of AD in TCM and syndrome differentiation and treatment, and provide strong support for cultivating excellent TCM clinical talents and improving the clinical efficacy of TCM. Summary of the invention
[0005] The purpose of the present invention is to provide an ADY urine biomarker and its application. The present invention uses non-targeted metabolomics technology to determine the difference in urine metabolites between ADY and ADNY patients, combines machine learning algorithms to build a diagnostic model, and determines the urine biomarker of Yin deficiency syndrome of Alzheimer's disease.
[0006] In order to achieve the above-mentioned object of the invention, the present invention provides the following technical solutions.
[0007] The invention discloses an early auxiliary diagnosis marker for Yin deficiency syndrome of Alzheimer's disease, which is characterized in that the marker is one or a combination of several of hesperidin, kaempferol, adrenaline, uridine monophosphate and aminobutyric acid.
[0008] Furthermore, the above-mentioned Alzheimer's disease Yin deficiency syndrome early auxiliary diagnosis marker is used in the preparation of Alzheimer's disease Yin deficiency syndrome early auxiliary diagnosis product.
[0009] The present invention also discloses the use of a substance for detecting the above-mentioned marker for early auxiliary diagnosis of Yin deficiency syndrome of Alzheimer's disease in the preparation of a product for early auxiliary diagnosis of Yin deficiency syndrome of Alzheimer's disease.
[0010] Furthermore, the substance is a substance used to detect the content of a diagnostic marker selected from urine.
[0011] Furthermore, the substance is an instrument and / or reagent required for an LC-MS / MS system combined quadrupole Orbitrap mass spectrometer for detecting early auxiliary diagnostic markers of Yin deficiency syndrome of Alzheimer's disease.
[0012] The present invention also discloses an early auxiliary diagnosis kit for Yin deficiency syndrome of Alzheimer's disease, which is characterized in that the kit contains substances for detecting the early auxiliary diagnosis markers for Yin deficiency syndrome of Alzheimer's disease mentioned above.
[0013] Furthermore, the kit also includes a carrier recording the model or a device loaded with the model.
[0014] The present invention also discloses a method for constructing a diagnosis model of Yin deficiency syndrome of Alzheimer's disease, which is characterized by comprising the following steps:
[0015] The five metabolites hesperidin, kaempferol, epinephrine, uridine monophosphate and aminobutyric acid, which had different contents between Yin deficiency syndrome and non-Yin deficiency syndrome of Alzheimer's disease, were determined by urine non-targeted metabolomics LC-MS detection. One or a combination of them was used as predictor factors, and five machine learning models were used to construct prediction models for the differential metabolite data of the discovery set.
[0016] Furthermore, the four machine learning models are Naive Bayes, Random Forest, Support Vector Machine, Xgboost, and Logistic Regression.
[0017] Compared with the prior art, the beneficial effects of the present invention are as follows.
[0018] The prior art does not have a standardized diagnostic technology for ADY, and the evaluation method is less studied, which seriously affects the development of clinical and scientific research work in traditional Chinese medicine. The present invention combines non-targeted metabolomics with machine learning algorithms for the first time for patients with Yin deficiency and non-Yin deficiency of Alzheimer's disease, determines urine biomarkers, and establishes a urine sample diagnostic model. The biomarker screening method of the present invention is highly operational, the model construction method is simple, the obtained diagnostic model has excellent effect, high sensitivity, and excellent specificity, and is suitable for the early diagnosis of ADY. The present invention can complete the diagnosis by simply taking urine for testing, without the need for additional tissue sample collection, which greatly reduces the risk of trauma. The model disclosed in the present invention is for ADY patients, the diagnostic process is simple and rapid, which is conducive to timely treatment, and has high clinical application and promotion value. BRIEF DESCRIPTION OF THE DRAWINGS
[0019] Figure 1 The PCA diagram of urine metabolites of ADY and ADNY. ADNY: AD non-Yin deficiency group; ADY: AD Yin deficiency group.
[0020] Figure 2 OPLS-DA diagram of urinary metabolites of ADY and ADNY. ADNY: AD non-Yin deficiency group; ADY: AD Yin deficiency group.
[0021] Figure 3 The volcano plots of ADY and ADNY urine metabolites. ADNY: AD non-Yin deficiency group; ADY: AD Yin deficiency group.
[0022] Figure 4 MetaboAnalyst was used to perform functional clustering for the differential metabolites of ADY and ADNY.
[0023] Figure 5 The ROC curve of the differential metabolites of ADY and ADNY. Among them, A is hesperidin; B is kaempferol; C is epinephrine; D is uridine monophosphate; E is aminobutyric acid
[0024] Figure 6 The ROC curves of five machine learning algorithms were compared based on the discovery set data.
[0025] Figure 7 ROC curve of the random forest algorithm on the validation set was used to screen out five markers.
[0026] Figure 8 The ROC curves of the five markers screened out on the validation set are shown below. A is hesperidin, B is kaempferol, C is epinephrine, D is uridine monophosphate, and E is aminobutyric acid. DETAILED DESCRIPTION
[0027] The present invention is further described in detail below with specific embodiments, but it should not be understood that the scope of the above subject matter of the present invention is limited to the following embodiments, and all technologies realized based on the content of the present invention belong to the scope of the present invention.
[0028] Unless otherwise specified, all reagents and materials used in the present invention are commercially available.
[0029] 1. Methods and reagents.
[0030] 1. Sample collection.
[0031] A total of 97 AD patients, including 58 ADY and 39 ADNY, were included as the discovery set. A total of 22 AD patients, including 11 ADY and 11 ADNY, were included as the validation set. Venous blood was drawn from the subjects in an inert separation gel coagulation tube in the fasting state, and urine samples were collected. After the blood collection tube was left to stand at 20℃~25℃ for 20~30min, it was centrifuged at 1800Xg at room temperature for 10min to obtain urine samples, which were stored at -80℃ for non-targeted metabolomics research.
[0032] 2. Quality control.
[0033] The QC before metabolomics injection is used to monitor the instrument status before injection and balance the chromatography-mass spectrometry system. The QC inserted in the middle of the sample test is used to evaluate the system stability during the entire experiment and perform sample correlation analysis. The QC after the sample test is completed is used for qualitative analysis of metabolites through secondary mass spectrometry. (Note: QC samples are made by mixing equal amounts of the samples to be tested, and are tested on the machine before, during, and after the LC-MS / MS injection of the samples to be tested.)
[0034] 3. Non-targeted metabolomics LC-MS detection of urine.
[0035] The LC-MS / MS system combined quadrupole Orbitrap mass spectrometer (Q Exactive Orbitrap, ThermoFisher Scientific, USA) was used. Data analysis was mainly divided into three parts: basic data analysis, advanced data analysis, and personalized data analysis. Basic data analysis was to perform univariate statistical analysis and multivariate statistical analysis (MVA) on the qualitative and quantitative results of the metabolome to screen for metabolites with significant differences; personalized data analysis was a series of bioinformatics analyses on metabolites with significant differences based on basic data analysis.
[0036] 4. Mass spectrometry conditions.
[0037] 4.1 Reagents
[0038]
[0039] 4.2 Instruments
[0040]
[0041] 4.3 Sample processing:
[0042] (1) Thaw the sample slowly on ice;
[0043] (2) Take 100 μL of sample and add 300 μL of methanol, vortex and mix thoroughly; sonicate at 4°C for 30 min;
[0044] (3) Incubate at -40 °C for 1 h, vortex for 30 s, and incubate at 4 °C for 0.5 h;
[0045] (4) Centrifugation at 12,000 rpm for 15 min at 4°C;
[0046] (5) Take all the supernatant and place it in a centrifuge tube at -40°C for 1 h;
[0047] (6) Centrifuge at 12,000 rpm for 15 min at 4°C;
[0048] (7) Take 200 μL of supernatant, add 5 μL of internal standard (0.5 mg / mL dichlorophenylalanine), mix well, and transfer to a sample injection vial.
[0049] 4.4 On-machine testing:
[0050] Instrument analysis platform: LC-MS (Waters, UPLC; Thermo, Q Exactive);
[0051] Chromatographic column: ACQUITY UPLC H SST3 (2.1*100mm1.8μm);
[0052] Chromatographic separation conditions: column temperature, 40°C; flow rate, 0.300 mL / min; mobile phase composition A: water (0.05% formic acid), B: acetonitrile; injection volume: 5 μL, automatic sampler temperature, 4°C.
[0053] The mobile phase gradient elution program is shown in Table 1.
[0054] Table 1. Mobile phase elution program.
[0055]
[0056] Mass spectrometry parameters.
[0057] Positive mode: heater temperature 300°C; sheath gas flow rate: 45arb; auxiliary gas flow rate: 15arb; tail gas flow rate: 1arb; electrospray voltage: 3.0KV; capillary temperature: 350°C; S-Lens RF Level, 30%.
[0058] Negative mode: heater temperature 300°C; sheath gas flow rate: 45arb; auxiliary gas flow rate: 15arb; tail gas flow rate: 1arb; electrospray voltage: 3.2KV; capillary temperature: 350°C; S-Lens RF Level, 60%.
[0059] Scanning mode: Full Scan (m / z 70-1050) and data-dependent secondary mass spectrometry (dd-MS2, TopN=10); Resolution: 70,000 (primary mass spectrometry) & 17,500 (secondary mass spectrometry). Collision mode: high energy collision dissociation (HCD).
[0060] 5. Statistical methods.
[0061] First, the original file (.raw) obtained by mass spectrometry detection was imported into Compound Discoverer3.1 (CD) software for spectrum processing and database search to obtain the qualitative and quantitative results of metabolites, and then the data was quality controlled to ensure the accuracy and reliability of the data results. Next, multivariate statistical analysis of metabolites was performed, including principal component analysis (PCA), orthogonal partial least squares discriminant analysis (OPLS-DA), etc., to reveal the differences in metabolites in different groups, and metabolite correlation analysis was used to reveal the relationship between samples and between metabolites.
[0062] 6. Construction of machine learning algorithms.
[0063] R (4.3.1) and RStudio (2023.09.0Build 463) were used, the mlr3verse (0.2.8) package was used for machine learning modeling, and the ggplot2 (3.4.3) package was used to visualize the results.
[0064] 2. Experimental results.
[0065] The baseline data of the sample patients were tested (Table 2). There was no statistical difference in the MMSE and MoCA scores between the two groups, indicating that there was no significant difference in cognitive function between ADY and ADNY. There was no statistical difference in gender and education between the two groups, indicating that the subsequent data did not require covariance correction.
[0066] Table 2 Urine metabolomics Patient baseline characteristics.
[0067]
[0068] The PCA graph shows that there are some differences between ADY and ADNY, but the distinction is not obvious. As a conventional dimensionality reduction analysis method, PCA only focuses on the differences in the metabolites themselves when reducing the dimensionality, but does not pay attention to the groups in which the samples are grouped. In order to further find out the differential metabolites between the two groups and find out the specific metabolites of ADY, we continue to use supervised OPLS-DA for dimensionality reduction analysis. The graph shows that the two groups can be clearly separated, indicating that there are differences in the urine metabolites of ADY and ADNY.
[0069] The differential metabolites were subjected to t-test, and the metabolites with P < 0.05 and Fold Change ≥ 2 or ≤ 0.5 were selected to draw a volcano plot. The results of the differential metabolites are shown in the following table.
[0070] Table 3 Differential metabolites of ADY and ADNY in urine. (Screening conditions P<0.05, Fold Change ≥2 or ≤0.5)
[0071]
[0072] In order to explore the metabolic functions involved in the above-mentioned differential metabolites, this study clustered the metabolite functions through the MetaboAnalyst 5.0 platform and found that the most relevant pathways were lactose synthesis, catecholamine biosynthesis, glutamate metabolism, pyrimidine metabolism, and tyrosine metabolism.
[0073] In order to evaluate the ability of a single differential metabolite to distinguish ADNY from ADY, this study drew the ROC curve. The highest average AUC was 0.771 for epinephrine.
[0074] Next, five machine learning models were used to model the differential metabolite data of the discovery set. The models used included naive Bayes, random forest, support vector machine, Xgboost, and logistic regression. The results showed that the AUC of the ROC curve of the random forest model was the highest, reaching 0.8295, which was the best model among the five algorithms. The above results also show that the model constructed by combining the five metabolites is better than the AUC of a single model. We independently verified the optimal random forest model on the validation set, and the results showed that the AUC of the ROC curve reached 0.6280, indicating that urine differential metabolites have a good distinguishing effect on ADNY and ADY. ROC analysis of the validation set of the five differential metabolites showed that the use of differential metabolites alone also had a good diagnostic effect on ADNY and ADY, up to 0.901. The AUCs from high to low were uridine monophosphate and aminobutyric acid, kaempferol, epinephrine, and hesperidin.
[0075] Results: PCA and OPLS-DA analysis showed that the composition of urine metabolites in the ADY and ADNY groups was significantly different. Combined with the FC value and P value, a total of 5 urine differential metabolites were screened out, namely hesperidin, kaempferol, epinephrine, uridine monophosphate and aminobutyric acid. By clustering the metabolite functions, it was found that the most relevant pathways were lactose synthesis, catecholamine biosynthesis, glutamate metabolism, pyrimidine metabolism, and tyrosine metabolism. The five differential metabolites were used to construct diagnostic models using five machine learning algorithms, including naive Bayes, random forest, support vector machine Xgboost and logistic regression. It was found that the random forest model had the highest AUC, which was also better than the AUC of a single differential metabolite, with an average AUC of 0.8295. The model was verified by an independent validation set, and the AUC reached 0.6280, indicating that urine metabolites have a good diagnostic effect on ADNY and ADY.
[0076] Conclusion: The urinary metabolite compositions of ADY and ADNY patients were significantly different. The random forest model constructed with hesperidin, kaempferol, epinephrine, uridine monophosphate and aminobutyric acid can effectively distinguish ADNY and ADY patients and can be used as ADY urine biomarkers.
[0077] The above description is only a preferred embodiment of the present invention and is not intended to limit the scope of the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.
Claims
1. An early auxiliary diagnostic marker for Yin deficiency syndrome of Alzheimer's disease, characterized by: The marker is one or a combination of hesperidin, kaempferol, epinephrine, uridine monophosphate and aminobutyric acid.
2. Use of the marker for early auxiliary diagnosis of Yin deficiency syndrome in Alzheimer's disease according to claim 1 in the preparation of a product for early auxiliary diagnosis of Yin deficiency syndrome in Alzheimer's disease.
3. Use of a substance for detecting the marker for early auxiliary diagnosis of Yin deficiency syndrome of Alzheimer's disease according to claim 1 in the preparation of a product for early auxiliary diagnosis of Yin deficiency syndrome of Alzheimer's disease.
4. The use according to claim 3, characterized in that: The substance is a substance used to detect the content of diagnostic markers selected from urine.
5. The use according to claim 4, characterized in that: The substances are instruments and / or reagents required for an LC-MS / MS system combined quadrupole Orbitrap mass spectrometer for detecting early auxiliary diagnostic markers of Yin deficiency syndrome of Alzheimer's disease.
6. An early auxiliary diagnosis kit for Yin deficiency syndrome of Alzheimer's disease, characterized by: The kit contains substances for detecting the early auxiliary diagnostic markers of Yin deficiency syndrome of Alzheimer's disease as described in claim 1.
7. The early auxiliary diagnosis kit for Yin deficiency syndrome of Alzheimer's disease according to claim 6, characterized in that: The kit also includes a carrier recording the model or a device loaded with the model.
8. A method for constructing a diagnostic model of Yin deficiency syndrome of Alzheimer's disease, characterized by: The following steps are involved: Using urine non-targeted metabolomics LC-MS detection to determine the content of five metabolites with different levels between Yin deficiency syndrome and non-Yin deficiency syndrome of Alzheimer's disease, one or a combination of hesperidin, kaempferol, epinephrine, uridine monophosphate and aminobutyric acid was used as predictor, and four machine learning models were used to construct prediction models for the differential metabolite data of the discovery set.
9. The construction method according to claim 8, characterized in that: The five machine learning models are Naive Bayes, Random Forest, Support Vector Machine, Xgboost and Logistic Regression.