Clinical decision support method for predicting smoke disease based on biomarkers in blood

Abstract

The invention discloses a clinical decision support method for identifying smoke disease based on biomarkers in blood. The method comprises the following steps: firstly, determining differential genes between a smoke disease patient and a healthy control group by analyzing a data set in a GEO database; biological pathways and disease types enriched by the differential genes are further analyzed; then identifying a biomarker related to the smoke disease through LASSO and SVM-RFE machine learning algorithms, and evaluating the predictive ability of the biomarker through an ROC curve; finally, screening to obtain CFD and DKFZp434L192 genes, and constructing a nomogram based on the expression of the two genes in the data set so as to predict smoke disease high-risk population in general population. Clinical samples are further collected, expression of CFD and DKFZp434L192 genes in whole blood of smoke disease patients is verified through RT-qPCR, and the technical problem that at present, no efficient recognizable cause of smoke disease exists is solved. By analyzing the expression of the characteristic gene in the blood of the patient, the high-risk group of the smoke disease is quickly predicted, and decision support is provided for timely intervention and treatment.

Description

Technical Field

[0001] The present invention relates to the field of biomedical technology, and in particular to a clinical decision support method for predicting moyamoya disease based on biomarkers CFD and DKFZp434L192 in blood. Background Art

[0002] Moyamoya disease (MMD), characterized by progressive stenosis of the terminal internal carotid artery (ICA) and an abnormal vascular network at the base of the brain, is a chronic, rare cerebrovascular occlusive disease with a poor prognosis and great regional and racial disparities. The occurrence and development of MMD is multifactorial, and its underlying etiology and pathogenic mechanisms are largely unknown.

[0003] The development of molecular biology and second-generation sequencing technology has made it possible to explore the pathogenesis of diseases on a large scale at the genetic and molecular levels. Therefore, the screening of biomarkers also provides a new direction for the prediction and treatment of diseases. However, there is relatively little attention paid to the screening of differentially expressed genes (DEGs) related to the occurrence and development of the disease through bioinformatics analysis of gene expression as a method for assessing the risk of moyamoya disease. Summary of the invention

[0004] In view of the problems of the prior art, the purpose of the present invention is to provide a clinical decision support method for predicting moyamoya disease based on the biomarkers CFD and DKFZp434L192 in the blood, so as to solve the problem that there is no efficient and identifiable biomarker for moyamoya disease. The present invention screens characteristic genes through machine learning, analyzes the expression of characteristic genes in the patient's blood, and quickly predicts the high-risk population for moyamoya disease, providing decision support for timely intervention and treatment.

[0005] To achieve the above object, the present invention adopts the following technical solutions:

[0006] The clinical decision support method for predicting moyamoya disease based on biomarkers in blood includes the following steps:

[0007] Using CFD and DKFZp434L192 genes as biomarkers, a nomogram based on the expression of biomarkers in the data set was constructed, and the risk probability of moyamoya disease in the general population was predicted by calculating the risk score (RiskScore); specifically: first, the differential genes between moyamoya disease patients and healthy controls were determined by analyzing the data set in the GEO database; then the biological pathways and disease types enriched by the differential genes were further analyzed; then the biomarkers related to moyamoya disease were identified by LASSO and SVM-RFE machine learning algorithms, and the predictive ability of the biomarkers was evaluated by the ROC curve; finally, the CFD and DKFZp434L192 genes were screened, and a nomogram was constructed based on the expression of the two genes in the data set, and the risk score (Risk Score) was calculated to predict the risk probability of moyamoya disease in the general population, so as to screen the high-risk population for moyamoya disease in the general population; further clinical samples were collected, and RT-qPCR was used to verify the expression of CFD and DKFZp434L192 genes in the whole blood of moyamoya disease patients.

[0008] Further, the CFD and DKFZp434L192 genes were screened out as candidate markers by the following steps:

[0009] S1, identification of differentially expressed genes between patients with MMD and healthy controls;

[0010] S2, further analysis of biological pathways and disease types enriched by differentially expressed genes;

[0011] S3, screening of biomarkers for predicting moyamoya disease;

[0012] S4, Validation of the differential expression of biomarkers in patients with MMD and healthy controls.

[0013] Preferably, in step S1, two moyamoya disease microarray data sets downloaded from the GEO database are analyzed to determine the differentially expressed genes between moyamoya disease patients and a healthy control group.

[0014] Preferably, in step S2, Metascape database, Gene Ontology and Kyoto Encyclopedia of Genes and Genomes enrichment analysis are used to further analyze the biological pathways and disease types enriched by the differentially expressed genes.

[0015] Preferably, in step S3, biomarkers of moyamoya disease patients are screened by LASSO and SVM-RFE machine learning algorithms, and the predictive ability of the biomarkers is evaluated by ROC curve to obtain biomarkers with higher predictive ability.

[0016] Preferably, in step S4, the expression of candidate biomarkers is further verified by RT-qPCR using external datasets and by collecting clinical blood samples to extract RNA.

[0017] Furthermore, the construction of the biomarker expression-based nomogram was based on the expression of CFD and DKFZp434L192 and was constructed using the "RMS" package.

[0018] Furthermore, the risk score is in the range of 0-1, and the closer the score is to 1, the higher the risk of moyamoya disease.

[0019] Furthermore, the biomarker expression-based nomogram is used to predict high-risk populations for moyamoya disease, providing rapid auxiliary decision support for the prevention and treatment of moyamoya disease.

[0020] Compared with the prior art, the present invention has the following advantages:

[0021] (1) The present invention analyzed two moyamoya disease microarray data sets downloaded from the GEO database, and determined the differentially expressed genes between moyamoya disease patients and healthy controls; further analyzed the biological pathways and disease types enriched by the differentially expressed genes by using the Metascape database, gene ontology, and Kyoto Encyclopedia of Genes and Genomes enrichment analysis; identified biomarkers related to moyamoya disease by using LASSO and SVM-RFE machine learning algorithms; and evaluated the predictive ability of the biomarkers by using ROC curves, and further verified the biomarkers by using data sets and RT-qPCR; finally, it was determined that the CFD and DKFZp434L192 genes could be used as biomarkers to predict the high-risk population for moyamoya disease, and the biomarkers showed good results in the prediction; a nomogram was constructed based on the expression of the two genes in the data set to predict the high-risk population for moyamoya disease in the general population, and the prediction effect was good, which solved the problem that there is currently no efficient and identifiable biomarker technology for moyamoya disease.

[0022] (2) The present invention uses machine learning to identify characteristic genes, with the goal of clinical decision-making, and further assists clinicians in providing rapid decision-making support for the prevention and treatment of moyamoya disease. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 Heat map (A) and volcano map (B) of differentially expressed genes in the patient group VS the control group of the combined related moyamoya disease datasets GSE157628 and GSE141025.

[0024] Figure 2 Results of differential gene enrichment analysis: A. GO enrichment analysis; B. DO enrichment analysis; C. KEGG enrichment analysis; D. Metascape analysis; E. PPI protein interaction network diagram.

[0025] Figure 3 Process diagram for screening biomarkers in the moyamoya disease dataset. AB. Lasso regression analysis; CD. SVM-RFE support vector machine recursive feature elimination analysis; E. Intersection of the two algorithms.

[0026] Figure 4 Process diagram for validating biomarkers in the validation set, A. Expression level of intersection genes in the combined training set B. ROC of intersection genes in the combined training set; C. Expression level of intersection genes in the validation set; D. ROC of intersection genes in the validation set.

[0027] Figure 5 The nomogram, decision curve, clinical impact curve, and calibration curve of candidate biomarkers (A: nomogram; B: decision curve; C: clinical impact curve; D: calibration curve).

[0028] Figure 6 This is the relative expression analysis of the target gene by qRT-PCR. DETAILED DESCRIPTION

[0029] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. 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.

[0030] Unless otherwise defined, all professional terms used hereinafter have the same meanings as those generally understood by those skilled in the art. The professional terms used herein are only for the purpose of describing specific embodiments and are not intended to limit the scope of protection of the present invention; unless otherwise specifically stated, various raw materials, reagents, instruments and equipment used in the present invention can be purchased from the market or prepared by existing methods.

[0031] Example 1

[0032] The GSE157628 and GSE141025 datasets were downloaded from the GEO database. Online tools were used to screen differentially expressed genes in moyamoya disease, and differentially expressed genes were identified when p value < 0.05 and |log2FC| > 1.

[0033] 1.1 Screening of Moyamoya disease-related datasets through the GEO database

[0034] The matrix files of datasets GSE157628 and GSE141025 are from the GEO database created and maintained by the National Center for Biotechnology Information (http: / / www.ncbi.nlm.nih.gov / geo / ). The GSE157628 dataset includes 11 samples from patients with moyamoya disease and 9 control samples. GSE141025 includes 4 samples from patients with moyamoya disease and 4 control samples. The probes in each dataset were converted to gene symbols according to their probe annotation files. For multiple probes corresponding to the same gene symbol, the probe average was calculated as the final expression value of the gene. The two datasets GSE157628 and GSE141025 were merged into a metadata queue for further integrated analysis. At the same time, the “SVA” package of R software was applied to eliminate batch effects. The limma package of R (http: / / www.bioconductor.org / ) was used for background correction, normalization between arrays, and differential expression analysis between moyamoya disease and control samples. |logFC|>1 and p<0.05 were defined as differential genes.

[0035] 1.2 Analysis of differential gene expression

[0036] First, hierarchical cluster analysis was performed on the 28 samples obtained to detect the correlation between samples. The limma package of R (http: / / www.bioconductor.org / ) was used to perform background correction, normalization between arrays, and differential expression analysis between samples and control samples. The screening conditions were set to |logFC|>1 and P<0.05, and the difference was statistically significant. The R language was used to draw heat maps and volcano maps for the differentially expressed genes. (e.g. Figure 1 ).

[0037] 1.3 Gene function annotation and enrichment analysis

[0038] In order to reveal the biological function of moyamoya disease, the “clusterProfiler”, “enrichplot”, “GSEABase” and other packages were used to perform functional and pathway enrichment analysis on differentially expressed genes (e.g. Figure 2 ).

[0039] Example 2

[0040] Screening and validation of landmark genes

[0041] The candidate diagnostic biomarkers were further screened by the least absolute shrinkage and selection operator model (LASSO), and the lasso regression algorithm was performed using the "glmnet" package in R. The support vector machine (SVM) model was constructed using the "e1071" package in R to screen the key genes. The intersection genes of the two were taken using a Venn diagram, and the differential genes of the intersection were further verified using the validation set GSE189993. The R package "pROC" was used to draw the ROC curve to verify the specificity and sensitivity of the candidate differential genes. Two machine learning algorithms were used to screen diagnostic biomarkers among 76 key genes. A total of 16 meaningful characteristic genes were identified using LASSO, such as Figure 3 A and Figure 3 B. Using the SVM algorithm, a total of 8 genes with significant characteristics were identified, such as Figure 3 C and Figure 3 D. The intersection of the two algorithms yields 6 overlapping genes, such as Figure 3 E. They were CFD, DKFZp434L192, EN1, MIA, MYOT and OGN, and their expressions in the MMD group were significantly different from those in the control group (P<0.001). Figure 4 A. Their AUCs were 0.873, 0.873, 0.839, 0.895, 0.898, and 0.930, respectively, indicating that the above six genes have the possibility of being used as diagnostic biomarkers for moyamoya disease, such as Figure 4 B. Subsequently, we used the dataset GSE189993 as the validation set and used the GSE189993 dataset to validate the above six diagnostic biomarkers. The expression results of the above six genes showed that CFD and DKFZp434L192 were significantly upregulated in the MMD group (P<0.05), such as Figure 4 C. ROC is an indicator for evaluating diagnostic accuracy and sensitivity. The ROC curve was used to test the diagnostic efficacy of the above six biomarkers. The results showed that the AUCs of CFD and DKFZp434L192 were 0.831 and 0.732, respectively, with high predictive values, indicating that the above two markers have good diagnostic value. Figure 4 D.

[0042] Example 3

[0043] Construction and evaluation of the prediction nomogram for moyamoya disease

[0044] Construction of a nomogram to predict the risk of moyamoya disease patients based on gene expression ( Figure 5 A). Decision curve analysis (DCA) shows that the “all influencing factors” curve is much higher than the gray line, which indicates that the line graph model is more accurate ( Figure 5B). In order to more intuitively evaluate the clinical effect of the nomogram model, a clinical impact curve based on the DCA curve was drawn. The high-risk curve represented by the thick solid line is very close to the true positive patient curve represented by the thick dotted line, which shows that the nomogram model has a good predictive ability ( Figure 5 C). The calibration curve showed that the nomogram model had a high accuracy in predicting the risk of moyamoya disease ( Figure 5 D). The calibration curve showed that the error between the risk of developing moyamoya disease and the predicted risk was small and statistically significant, indicating that the nomogram model had a high accuracy in predicting moyamoya disease.

[0045] The present invention collected blood samples from normal subjects and patients with moyamoya disease, and detected the mRNA expression levels of CFD and DKFZp434L192 by RT-qPCR. It was found that the expression levels of the two markers could be detected by RT-qPCR at the mRNA level ( Figure 6 ) and brought into the model as a diagnostic marker for moyamoya disease.

[0046] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.

Claims

1. A clinical decision support method for predicting moyamoya disease based on biomarkers in blood, characterized in that: The following steps are involved: Using CFD and DKFZp434L192 genes as biomarkers, a nomogram based on biomarker expression was established; the risk probability of moyamoya disease in the general population was predicted by calculating the risk score.

2. The clinical decision support method for predicting moyamoya disease based on blood biomarkers according to claim 1, characterized in that: The CFD and DKFZp434L192 genes are selected as candidate markers by the following steps: S1, identification of differentially expressed genes between patients with MMD and healthy controls; S2, further analysis of biological pathways and disease types enriched by differentially expressed genes; S3, screening of biomarkers for predicting moyamoya disease; S4, Validation of the differential expression of biomarkers in patients with MMD and healthy controls.

3. The clinical decision support method for predicting moyamoya disease based on blood biomarkers according to claim 2, characterized in that: In the step S1, two moyamoya disease microarray data sets downloaded from the GEO database are analyzed to determine the differentially expressed genes between moyamoya disease patients and a healthy control group.

4. The clinical decision support method for predicting moyamoya disease based on blood biomarkers according to claim 2, characterized in that: In step S2, Metascape database, Gene Ontology and Kyoto Encyclopedia of Genes and Genomes enrichment analysis are used to further analyze the biological pathways and disease types enriched by the differentially expressed genes.

5. The clinical decision support method for predicting moyamoya disease based on blood biomarkers according to claim 2, characterized in that: In step S3, biomarkers of moyamoya disease patients are screened by LASSO and SVM-RFE machine learning algorithms, and the predictive ability of biomarkers is evaluated by ROC curve.

6. The clinical decision support method for predicting moyamoya disease based on blood biomarkers according to claim 2, characterized in that: In step S4, the candidate biomarker expression is further validated by using external datasets and by collecting clinical blood samples and extracting RNA followed by RT-qPCR.

7. The clinical decision support method for predicting moyamoya disease based on blood biomarkers according to claim 1, characterized in that: The nomogram was constructed based on the expression of CFD and DKFZp434L192 in the dataset using the "RMS" package.

8. The clinical decision support method for predicting moyamoya disease based on blood biomarkers according to claim 1, characterized in that: The risk score ranges from 0 to 1, and the closer the score is to 1, the higher the risk of developing moyamoya disease.

9. The clinical decision support method for predicting moyamoya disease based on blood biomarkers according to claim 1, characterized in that: The nomogram is used to predict the high-risk population for moyamoya disease and provide rapid auxiliary decision support for the prevention and treatment of moyamoya disease.

Images