Alzheimer's disease biomarker and screening method thereof
By integrating differential expression analysis, protein-protein interaction networks, machine learning, and Mendelian randomized causal inference, the CD44 gene was selected as a biomarker for Alzheimer's disease, solving the problems of lack of early diagnostic tools and complex detection, and achieving efficient and accurate early diagnosis of Alzheimer's disease.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-31
- Publication Date
- 2026-04-10
AI Technical Summary
Existing early diagnostic tools for Alzheimer's disease are lacking, the tests are complex and have low accuracy, blood biomarkers have insufficient sensitivity and limited specificity, traditional methods lack multi-level and multi-dimensional data integration and analysis, resulting in a high false positive rate, and pathological changes in the brain cannot be stably detected in peripheral blood.
By screening differentially expressed genes related to ferroptosis in the GEO database, performing protein-protein interaction network analysis, and combining machine learning and Mendelian randomized causal inference, the CD44 gene was selected as a biomarker for Alzheimer's disease, and verified by human plasma RT-qPCR.
The CD44 gene was successfully screened as a key biomarker for Alzheimer's disease, which improved the reliability and specificity of early diagnosis, provided a non-invasive or minimally invasive detection method, enhanced the understanding of the pathogenesis of the disease, and provided a new direction for intervention strategies.
Smart Images
Figure CN121838864A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of biomarker technology, and more specifically, to an Alzheimer's disease biomarker and a screening method thereof. Background Technology
[0002] Alzheimer's disease (AD) is an age-related, progressive neurodegenerative disease, clinically characterized by memory impairment, cognitive decline, and behavioral abnormalities. It severely impacts patients' quality of life and places a heavy burden on society. Currently, the clinical diagnosis of AD primarily relies on neuropsychological assessments and brain imaging examinations. However, these methods often only provide a definitive diagnosis when the disease has progressed to the middle or late stages, lacking early, sensitive, and specific biomarkers.
[0003] While cerebrospinal fluid (CSF)-based biomarkers (such as Aβ42 and tau protein) have some diagnostic value, their invasive collection process, complex operation, and low patient acceptance limit their widespread clinical application. In recent years, research on peripheral blood (such as plasma) biomarkers has gained increasing attention due to their convenient sampling and minimal invasiveness, making them more suitable for early screening and dynamic monitoring. However, existing blood biomarkers still suffer from insufficient sensitivity, limited specificity, and poor reproducibility. Furthermore, traditional views suggest that pathological changes in the brain may not be reliably detectable in peripheral blood. In addition, traditional biomarker screening methods often rely on single-omics data analysis (such as transcriptomics or proteomics), lacking multi-level and multi-dimensional integrated analysis, which often leads to high false-positive rates and weak clinical interpretability of the selected biomarkers.
[0004] Therefore, there is an urgent need in this field for a systematic approach that can efficiently and accurately screen out blood biomarkers for Alzheimer's disease (AD) from massive amounts of data that are supported by causal evidence and are easy to verify in clinical samples, in order to promote the early diagnosis and precise intervention of AD. Summary of the Invention
[0005] Alzheimer's disease (AD) is a progressive neurodegenerative disease. Ferroplasmosis promotes disease progression through oxidative stress and other mechanisms, but the role of ferroptosis-related genes in its pathogenesis is not fully understood. To address these technical problems, the present invention aims to provide an Alzheimer's disease biomarker and its screening method, thereby resolving the current issues of a lack of early diagnostic tools, complex testing methods, and low accuracy in Alzheimer's disease diagnosis.
[0006] The objective of this invention is achieved through the following technical solution: In a first aspect, the present invention provides a method for screening biomarkers for Alzheimer's disease, comprising the following steps: S1. Screen for differentially expressed genes in AD patients using the GEO database and obtain differentially expressed genes in AD related to ferroptosis; S2. Protein-protein interaction network analysis of differentially expressed genes in AD related to ferroptosis; S3. Using machine learning, preliminary key feature genes were screened from differentially expressed genes associated with ferroptosis (AD). S4. Perform two-sample Mendelian randomization analysis on the selected characteristic genes to screen out key characteristic genes with potential causal association with AD. S5. Perform human plasma RT-qPCR verification on the key characteristic genes of causal association that have potential causal association with AD screened in step S4, and identify the genes that show statistically significant differences in expression levels compared with the control group in the verification as the biomarkers of Alzheimer's disease.
[0007] As some specific embodiments of the present invention, step S1 specifically includes: S11. Obtain gene expression datasets of AD patients and healthy controls from the GEO database; S12. Use the limma package in R language to perform differential expression analysis on gene expression data of AD patients and healthy control samples, and obtain differentially expressed genes in AD patients; S13. Obtain ferroptosis-related genes from the FerrDb database; S14. The intersection of the ferroptosis-related genes and the differentially expressed genes of AD patients is used to obtain the ferroptosis-related AD differentially expressed genes.
[0008] As some specific embodiments of the present invention, in step S2, the protein interaction network is constructed using the STRING database, and visualized and analyzed using Cytoscape software.
[0009] As some specific embodiments of the present invention, in step S3, preliminary key feature genes are screened from differentially expressed genes of AD related to ferroptosis by combining the machine learning algorithms LASSO regression and Random Forest.
[0010] Furthermore, step S3 specifically includes the following steps: S31. Using the glmnet package in R language, perform LASSO regression analysis on the differentially expressed genes of ferroptosis-related AD, determine the best model and screen characteristic genes through cross-validation; S32. Use the Random Forest package in R language to perform random forest analysis on the differentially expressed genes of AD related to ferroptosis, and screen characteristic genes by ranking them according to the importance of the Gini index; S33. Take the intersection of the screening results from the two methods to obtain the preliminary key feature genes; S34. Use ROC curves to evaluate the diagnostic performance of the preliminary key characteristic genes.
[0011] As some specific embodiments of the present invention, step S4 specifically includes the following steps: S41. Obtain GWAS data of AD and eQTL data of the preliminary key feature genes from the IEU OpenGWAS database; S42. Select SNPs that meet the preset conditions as instrumental variables; S43. For each preliminary key feature gene, use the "TwoSampleMR" package in R language, adopt the inverse variance weighting method as the main analysis method to perform causal inference, evaluate the causal effect of the gene on AD, and use the MR-Egger method and weighted median estimation method as supplementary analysis methods. S44. Sensitivity analysis was performed using leave-one-out method and MR-Egger regression to verify the robustness of the causal inference results. S45. Genes with a preliminary key feature value less than 0.05 in the inverse variance weighted analysis are identified as key feature genes with a potential causal relationship with AD.
[0012] As some specific embodiments of the present invention, step S5 specifically includes the following steps: S51. Total RNA was extracted from plasma samples of AD patients and healthy controls and reverse transcribed into cDNA; S52. Quantitative PCR amplification is performed using specific primers for the key characteristic genes of causal associations screened in step S4. S53. Using the GAPDH gene as an internal reference gene, the relative expression levels of the key characteristic genes of the causal association were calculated using the ΔΔCT method. S54. Perform statistical analysis to verify the difference in the expression level of the key characteristic gene of the causal association between the AD patient group and the control group. If the expression level of the gene in the AD patient group is statistically significantly different from that in the healthy control group (P<0.05), then the gene is finally identified as the Alzheimer's disease biomarker.
[0013] Secondly, the present invention provides a biomarker for Alzheimer's disease, wherein the biomarker is the CD44 gene or the protein encoded therein, obtained by screening using any of the screening methods described above.
[0014] Thirdly, the present invention provides a kit for detecting Alzheimer's disease, comprising reagents for specifically detecting CD44 gene expression levels.
[0015] As some specific embodiments of the present invention, the reagent includes primer pairs for specifically amplifying the CD44 gene, the sequences of which are shown in SEQ ID NO.1 and SEQ ID NO.2.
[0016] As some specific embodiments of the present invention, the kit is used for testing human plasma samples.
[0017] Fourthly, the present invention provides the application of the Alzheimer's disease biomarker or kit in the preparation of products for diagnosing or assisting in the diagnosis of Alzheimer's disease.
[0018] Compared with the prior art, the present invention has the following beneficial effects: (1) This invention provides a biomarker for Alzheimer's disease and a screening method thereof. For the first time, a systematic screening strategy integrating differential expression analysis, protein interaction networks, machine learning, Mendelian randomized causal inference and clinical sample validation was successfully used to screen CD44 as a key biomarker for AD, which is crucial for understanding the molecular mechanism of Alzheimer's disease and exploring precision treatment.
[0019] (2) This invention integrates two machine learning algorithms to screen characteristic genes with high predictive performance and combines them with Mendelian randomization to explore genes causally related to Alzheimer's disease (AD) as biomarkers. To ensure the reliability of the research results in clinical practice, clinical samples are used for experimental verification. Compared with traditional correlation studies that only reveal associations, Mendelian randomization promotes a deeper study of causality by using genetic variation, which effectively reduces the influence of confounding variables and improves the internal validity of the research results. This invention combines qPCR technology and uses bioinformatics analysis methods to screen key genes related to ferroptosis in peripheral blood of Alzheimer's patients and conducts experimental verification at the clinical specimen level. This breaks through the limitations of traditional methods that can only reveal associations. Mendelian randomization provides causal evidence at the genetic level, significantly improving the reliability and biological explanatory power of biomarker discovery.
[0020] (3) The marker CD44 identified in this invention is expressed in peripheral blood, the sample is easy to obtain, and the trauma is small. It overcomes the bottleneck of the difficulty in clinical promotion of cerebrospinal fluid markers and provides a new feasible target for non-invasive or minimally invasive early diagnosis of AD.
[0021] (4) This invention links ferroptosis, a specific cell death mechanism, with the molecular pathology of AD, providing new clues and directions for understanding the pathogenesis of AD and developing novel intervention strategies targeting the ferroptosis pathway.
[0022] (5) This invention uses CD44 as a biomarker for Alzheimer's disease, which has excellent predictive effect, provides new clues for the diagnosis and treatment of AD, and lays the foundation for revealing its complex pathogenesis and developing more effective intervention measures. Attached Figure Description
[0023] Other features, objects, and advantages of the present invention will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings: Figure 1 This is a volcano diagram of differentially expressed genes in AD patients in Example 1; Figure 2 The Venn diagram is the intersection of ferroptosis-related genes and differentially expressed genes in AD in Example 1; Figure 3 This is a protein-protein interaction network diagram of differentially expressed genes related to ferroptosis AD constructed based on the STRING database in Example 1; Figure 4 This is a ranking of the importance of characteristic genes screened by random forest analysis in Example 1; Figure 5 This is the ROC curve of the diagnostic efficacy of the key feature genes in Example 1; Figure 6 This is a Mendelian randomized scatter plot of the causal effect of CD44 on AD in Example 1; Figure 7 This is a forest diagram illustrating the causal effects of each SNP on individuals at risk of AD, as shown in Example 1. Figure 8 This is a leave-one-out forest plot of the causal effect of CD44 on AD in Example 1; Figure 9 This is a bar chart showing the relative expression levels of CD44 in plasma between the AD patient group and the healthy control group in Example 1. Detailed Implementation
[0024] The present invention will now be described in detail with reference to specific embodiments. These embodiments will help those skilled in the art to further understand the present invention, but do not limit the invention in any way. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention. These all fall within the scope of protection of the present invention.
[0025] It should be noted that the scientific and technical terms and their abbreviations used in this invention have meanings commonly understood by those skilled in the art. The following is a list of some of the terms and abbreviations used in this invention: p-value: Represents the p-value in significance analysis. The smaller the p-value, the more significant the difference. AD: Alzheimer's disease; MR: Mendelian randomization; IVW: inverse variance weighted method; MR-PRESSO: MR Pleiotropy RESidual Sum Outlier, a Mendelian randomized pleiotropy residual and outlier detection method. RT-qPCR: reverse transcription-quantitative polymerase chain reaction; ROC: receiver operating characteristic; FRGs: Ferroptosis-related genes; DEGs: Differentially expressed genes, genes differentially expressed in AD patients; DEFRGs: Differentially expressed ferroptosis-related genes; PPI: protein-protein interaction; gene / protein interaction. Example 1: Screening of biomarkers for Alzheimer's disease 1.1 Data Sources and Screening of Differentially Expressed Genes A dataset of AD-related gene expression (e.g., GSE122063) was obtained from the GEO database (https: / / www.ncbi.nlm.nih.gov / geo / ), containing AD patient samples (n=92) and healthy control samples (n=44). Differential expression analysis was performed on the samples using R (version 4.2.1) and the limma package. The screening criteria were set as |log2 Fold Change |>1 and adj.P<0.05, resulting in 312 differentially expressed genes (DEGs) from AD patients. These are shown in Table 1 below. Table 1 Differentially expressed genes (DEGs) in AD patients
[0026] like Figure 1 As shown, this is a volcano plot of 312 differentially expressed genes (DEGs) in Table 1 above, of which 139 genes are upregulated (red) and 173 genes are downregulated (green).
[0027] 1.2 Acquisition of differentially expressed genes related to ferroptosis in Alzheimer's disease (AD) 484 ferroptosis-related genes (FRGs) were collected from the FerrDb database (http: / / www.zhounan.org / ferrdb / ). The intersection of these 484 FRGs with the 312 DEGs obtained in step 1.1 was used to construct a Venn diagram, as shown below. Figure 2 As shown, eight differentially expressed genes (DEFRGs) associated with ferroptosis were obtained, including CD44, CIRBP, HMOX1, HSPB1, NEAT1, NUPR1, PEX6, and PROM2.
[0028] 1.3 Protein-protein interaction network analysis The eight differentially expressed AD genes (DEFRGs) related to ferroptosis obtained in step 1.3 were imported into the STRING database (https: / / cn.string-db.org / ), the species was set to "Homo sapiens", and the minimum interaction score was set to 0.4 to obtain protein-protein interaction (PPI) network data.
[0029] Visualize the PPI network using Cytoscape 3.10.3 software, such as... Figure 3 As shown, the core modules in the network were analyzed using the MCODE plugin, and the top 10 genes with the highest connectivity were screened as candidate key genes. The top 10 candidate key genes with the highest connectivity include: CCL2, BDNF, TLR2, CD163, CD44, CD86, SST, TAC1, CCR5, and SPP1. Network analysis indicates that these candidate key genes may play a crucial regulatory role in the pathogenesis of ferroptosis-related AD.
[0030] 1.4 Machine Learning for Feature Gene Selection 1.4.1 LASSO Regression Analysis: DEFRGs were analyzed using the glmnet package in R language. The cross-validation fold was set to 10, and the "family" parameter was set to "binomial". The optimal lambda value (lambda.1se) was determined through 10-fold cross-validation, and the characteristic genes under this model were screened out, including: CD44, CIRBP, HMOX1, HSPB1, NEAT1, NUPR1, PEX6, and PROM2.
[0031] 1.4.2 Random Forest Analysis: A model was constructed using the Random Forest package in R, with 500 training trees. Gene importance was ranked, and genes with a Gini index > 0.5 were selected. Figure 4 As shown.
[0032] The gene importance ranking is as follows: CIRBP, HSPB1, PEX6, HMOX1, NUPR1, NEAT1, PROM2, CD44.
[0033] 1.4.3 Characteristic gene determination: The intersection of the screening results of the two methods mentioned above is used to obtain the key characteristic genes.
[0034] Key characteristic genes include: CD44, CIRBP, HMOX1, HSPB1, NEAT1, NUPR1, PEX6, and PROM2.
[0035] Use the pROC package to plot ROC curves, such as... Figure 5 As shown, the AUC value is calculated. The AUC value ranges from 0.5 to 1.0. The closer the value is to 1, the better the diagnostic performance.
[0036] 1.5 Mendelian Randomization Analysis 1.5.1 Data Preparation: Obtain AD-related genome-wide association study (GWAS) data and characteristic gene-related SNP data identified in step 1.4.3 from the IEU OpenGWAS database (https: / / gwas.mrcieu.ac.uk / ). Select SNPs that meet the following criteria as instrumental variables: ① P < 5 × 10⁻⁶. -8 ② Minimum allele frequency > 0.01; ③ Linkage disequilibrium r² < 0.001, distance > 1000 kb. The following SNPs were selected: rs1375493, rs138768880, rs507230, rs79713903, rs191739259, rs927335, rs6484785. To avoid false positives and bias, the palindromic sequence rs191739259 was deleted.
[0037] 1.5.2 Causal Relationship Analysis: Using the R language packages "TwoSampleMR" and "gwasglue", the IVW method was used as the main analysis method, with MR-Egger and weighted median methods used as supplementary analyses. A causal relationship was considered to exist when P < 0.05. The results are shown in Table 2 below.
[0038] Table 2 Causal Effects of CD44 and AD
[0039] like Figure 6The figure shows a scatter plot of the causal effect of CD44 on AD. like Figure 7 As shown, this is a forest plot illustrating the causal effect of each single nucleotide polymorphism on the risk of AD.
[0040] Mendelian randomization analysis showed that there was a statistically significant causal association between elevated CD44 gene expression and increased risk of Alzheimer's disease (AD), providing genetic causal evidence for CD44 as a blood biomarker for AD.
[0041] 1.5.3 Sensitivity analysis: The stability of instrumental variables was assessed by leave-one-out method and MR-Egger regression to exclude potential outliers and level pleiotropy.
[0042] The results of the heterogeneity test are shown in Table 3 below: Table 3 Heterogeneity Test
[0043] The results of the pleiotropy test are shown in Table 4 below: Table 4. Multivariability Test
[0044] like Figure 8 As shown, this is a leave-one forest diagram of the causal effect of CD44 on AD.
[0045] Sensitivity analysis results indicate that the causal inferences in this Mendelian randomization study exhibit good robustness, specifically including: (1) In the heterogeneity test (Table 3), the Q_pval of both the MR Egger method and the IVW method is greater than 0.05, indicating that there is no significant heterogeneity in the effect of each instrumental variable (SNP) on the outcome, and the estimation results among the instrumental variables are relatively consistent.
[0046] (2) The Egger intercept term pval in the pleiotropy test (Table 4) is greater than 0.05, indicating that there is no significant level pleiotropy, that is, the SNP mainly affects the outcome (AD) by influencing the exposure (CD44) rather than through other confounding pathways, which supports the instrumental variables satisfying the independence and exclusivity assumptions.
[0047] (3) Leave-one-out method refers to removing SNPs one by one and then performing Mendelian randomization analysis. Each point in the leave-one-out forest plot represents a SNP, used to observe the offset of each SNP on the horizontal axis to see if there are outliers. Leave-one-out forest plot ( Figure 8 The left-one-out forest diagram represents the causal effect of CD44 on the risk of AD when one SNP is omitted. Figure 8 The results show that there are no SNPs that could change the causal effect estimate, further demonstrating the robustness of the MR analysis results.
[0048] In summary, the sensitivity analysis results collectively confirm that the instrumental variables used in this study are reasonable, and the conclusion of "a positive causal association between CD44 and AD" derived from Mendelian randomized principal analysis (IVW) is robust and reliable, further strengthening the causal evidence for CD44 as a risk biomarker for AD.
[0049] 1.6 Human plasma RT-qPCR validation 1.6.1 Sample collection: Plasma samples from AD patients (n=3) and healthy controls (n=3) were collected. All samples were approved by the ethics committee and the subjects signed informed consent forms.
[0050] 1.6.2 RNA extraction and reverse transcription: (1) mRNA was extracted from plasma using RNA extraction buffer (G3013, Wuhan Servicebio Technology Co., Ltd, China); (2) The concentration and purity of the extracted mRNA were determined using an ultra-micro spectrophotometer (NanoDrop2000, Thermo, USA); (3) Stem-loop reverse transcription was performed using the SweScript RT I First Strand cDNA Synthesis Kit (G3330, Wuhan Servicebio Technology Co., Ltd, China); (4) The relative expression level of the target gene was assessed using 2×Universal Blue SYBR Green qPCR Master Mix (G3326, Wuhan Servicebio Technology Co., Ltd, China); (5) Amplification was performed using a real-time PCR instrument (CFX Connect, Bio-rad, USA).
[0051] In step (2) above, the concentration and purity of the extracted mRNA were measured as shown in Table 5 below: Table 5. mRNA Concentration and Purity
[0052] In step (4) above, the relative expression levels of the target gene are shown in Table 6 below: Table 6. Relative expression levels of target genes
[0053] In step (4), the designed qPCR primers are shown in Table 7 below: Table 7. qPCR primer sequence list
[0054] 1.6.4 Results Analysis: The relative expression level of CD44 was calculated using the ΔΔCT method, and the results are as follows: Figure 9 As shown, the expression level of CD44 in the AD patient group was significantly higher than that in the control group (P<0.05), confirming that CD44 can serve as a biomarker for AD.
[0055] The specific embodiments of the present invention have been described above. It should be understood that the present invention is not limited to the specific embodiments described above, and those skilled in the art can make various modifications or variations within the scope of the claims, which do not affect the essence of the present invention.
Claims
1. A method for screening biomarkers for Alzheimer's disease, characterized in that, Includes the following steps: S1. Screen for differentially expressed genes in AD patients using the GEO database and obtain differentially expressed genes in AD related to ferroptosis; S2. Protein-protein interaction network analysis of differentially expressed genes in AD related to ferroptosis; S3. Using machine learning, preliminary key feature genes were screened from differentially expressed genes associated with ferroptosis (AD). S4. Perform two-sample Mendelian randomization analysis on the preliminarily selected key feature genes to screen out key feature genes with potential causal association with AD. S5. Perform human plasma RT-qPCR verification on the key characteristic genes of causal association selected in step S4, and identify the genes that show statistically significant differences in expression levels compared with the control group during the verification as biomarkers of Alzheimer's disease.
2. The screening method according to claim 1, characterized in that, Step S1 specifically includes the following steps: S11. Obtain gene expression datasets of AD patients and healthy controls from the GEO database; S12. Use the limma package in R language to perform differential expression analysis on gene expression data of AD patients and healthy control samples, and obtain differentially expressed genes in AD patients; S13. Obtain ferroptosis-related genes from the FerrDb database; S14. The intersection of the ferroptosis-related genes and the differentially expressed genes of AD patients is used to obtain the ferroptosis-related AD differentially expressed genes.
3. The screening method according to claim 1, characterized in that, In step S2, the protein interaction network is constructed using the STRING database, and visualized and analyzed using Cytoscape software.
4. The screening method according to claim 1, characterized in that, In step S3, preliminary key feature genes are screened from differentially expressed genes in AD related to ferroptosis by combining the machine learning algorithms LASSO regression and Random Forest. Step S3 specifically includes the following steps: S31. Using the glmnet package in R language, perform LASSO regression analysis on the differentially expressed genes of ferroptosis-related AD, determine the best model and screen characteristic genes through cross-validation; S32. Use the Random Forest package in R language to perform random forest analysis on the differentially expressed genes of AD related to ferroptosis, and screen characteristic genes by ranking them according to the importance of the Gini index; S33. Take the intersection of the screening results from the two methods to obtain the preliminary key feature genes; S34. Use ROC curves to evaluate the diagnostic performance of the preliminary key characteristic genes.
5. The screening method according to claim 1, characterized in that, Step S4 specifically includes the following steps: S41. Obtain GWAS data of AD and eQTL data of the preliminary key feature genes from the IEU OpenGWAS database; S42. Select SNPs that meet the preset conditions as instrumental variables; S43. For each preliminary key feature gene, use the "TwoSampleMR" package in R language, adopt the inverse variance weighting method as the main analysis method to perform causal inference, evaluate the causal effect of the gene on AD, and use the MR-Egger method and weighted median estimation method as supplementary analysis methods. S44. Sensitivity analysis was performed using leave-one-out method and MR-Egger regression to verify the robustness of the causal inference results. S45. Genes with a preliminary key feature value of less than 0.05 in the inverse variance weighted analysis are identified as key feature genes with potential causal relationship with AD.
6. The screening method according to claim 1, characterized in that, Step S5 specifically includes the following steps: S51. Total RNA was extracted from plasma samples of AD patients and healthy controls and reverse transcribed into cDNA; S52. Quantitative PCR amplification is performed using specific primers for the key characteristic genes of causal associations screened in step S4. S53. Using the GAPDH gene as an internal reference gene, the relative expression levels of the key characteristic genes of the causal association were calculated using the ΔΔCT method. S54. Perform statistical analysis to verify the difference in the expression levels of the key characteristic genes of the causal association between the AD patient group and the control group. If the expression level of the gene in the AD patient group is statistically significantly different from that in the healthy control group, then the gene is finally identified as the Alzheimer's disease biomarker.
7. A biomarker for Alzheimer's disease, characterized in that, The biomarker is the CD44 gene or the protein it encodes, obtained by screening using the screening method described in any one of claims 1-6.
8. A reagent kit for detecting Alzheimer's disease, characterized in that, It includes reagents for specifically detecting the expression levels of the Alzheimer's disease biomarkers as described in claim 7.
9. The reagent kit according to claim 8, characterized in that, The reagent includes primer pairs for specifically amplifying the CD44 gene, the sequences of which are shown in SEQ ID NO.1 and SEQ ID NO.2; And / or, the kit is used to test human plasma samples.
10. The use of an Alzheimer's disease biomarker as described in claim 7 or a kit as described in claim 8 or 9 in the preparation of a product for the diagnosis or auxiliary diagnosis of Alzheimer's disease.