Biomarker for diagnosing osteoporosis with atherosclerosis as well as construction and application of column diagram

Abstract

The invention belongs to the technical field of medical biomolecules, and particularly discloses a column diagram construction method for osteoporosis accompanied with atherosclerosis, which specifically comprises the following steps: acquiring a GSE1428 data set, a GSE230665 data set and a GSE1000927 data set from a GEO database; screening out differential expression genes between the osteoporosis and normal tissues and between the atherosclerosis and healthy tissues by using a Limma bag; then performing WGCNA analysis on DEGs, and finally screening differential expression genes by using an LASSO regression algorithm, a random forest and an SVM-RFE algorithm to obtain a potential key common biomarker; based on the key common biomarkers, a column diagram is constructed, an ROC curve is drawn, and the diagnostic ability of candidate common biomarker genes is evaluated to reach 0.99. Based on the column diagram, a prediction system is provided, and a visual basis is provided for clinical diagnosis of osteoporosis with atherosclerosis.

Description

Technical Field

[0001] The invention belongs to the technical field of medical biomolecules, and relates to the construction of a biomarker and a nomogram for diagnosing osteoporosis with atherosclerosis, and also relates to the application of the nomogram. Background Art

[0002] Osteoporosis is a disease characterized by decreased bone density and destruction of bone tissue microstructure, which leads to increased bone brittleness and, in turn, an increased risk of fractures. According to the definition of the World Health Organization (WHO), osteoporosis refers to a state of increased risk of fractures due to decreased bone mass and decreased bone tissue quality. The disease is often called an "invisible disease" because in the early stages, patients often have no obvious symptoms and are not diagnosed until a fracture occurs. The incidence of osteoporosis is particularly high in women, especially after menopause, which is closely related to the decline in estrogen levels. Fractures not only affect the quality of life of patients, but can also lead to long-term physical dysfunction and psychological problems, increasing medical costs and social burdens.

[0003] Atherosclerosis is an arteriosclerotic disease caused by lipid deposition in blood vessels, inflammatory response and endothelial damage. As the disease progresses, the arterial wall gradually thickens and loses elasticity, and the vascular lumen narrows, which may eventually lead to serious cardiovascular and cerebrovascular events such as myocardial infarction and stroke. The occurrence of atherosclerosis is related to a variety of factors, including lifestyle factors such as hypertension, hyperlipidemia, diabetes and smoking. The disease is particularly common in the elderly and is an important cause of cardiovascular and cerebrovascular diseases in the elderly population.

[0004] In some elderly groups, osteoporosis is often accompanied by atherosclerosis. Traditional diagnostic methods may only be based on clinical symptoms or imaging examinations, which have certain limitations. Early diagnosis and intervention are crucial to improving the prognosis of patients. Although conventional imaging examinations and biochemical index monitoring can assist in diagnosis, they often have problems of poor timeliness and insufficient accuracy. Therefore, identifying new biomarkers and establishing corresponding nomograms, especially in the early stages, can help clinicians more accurately assess the risk of osteoporosis with atherosclerosis and provide patients with personalized treatment plans, thereby improving the efficiency and effectiveness of disease management. Summary of the invention

[0005] In view of this, the purpose of the present invention is to provide a diagnostic biomarker for atherosclerosis associated with osteoporosis and the construction of a nomogram, and also to a system for predicting or diagnosing osteoporosis accompanied by atherosclerosis based on the nomogram.

[0006] In order to achieve the above object, the present invention provides the following technical solutions:

[0007] 1. A gene marker for predicting or diagnosing osteoporosis accompanied by atherosclerosis, wherein the gene marker is COL1A1, SMAD2, PTEN or / and FBN1.

[0008] 2. Use of a reagent for detecting the expression level of a gene marker in the preparation of a product for predicting or diagnosing osteoporosis associated with atherosclerosis, wherein the gene marker is COL1A1, SMAD2, PTEN or / and FBN1.

[0009] 3. A method for constructing a nomogram for osteoporosis associated with atherosclerosis, comprising the following steps:

[0010] 1) Obtain GSE1428, GSE230665 and GSE1000927 datasets from the GEO database;

[0011] 2) Using the Limma package, we set |log2 fold change (FC)|>1.5 and P<0.05 to screen the differentially expressed genes between osteoporosis and normal tissues, and between atherosclerosis and healthy tissues;

[0012] 3) Identify candidate gene modules for atherosclerosis using weighted gene co-expression network analysis based on Pearson correlation;

[0013] 4) Use the STRING database to construct a protein-protein interaction network and perform functional enrichment analysis;

[0014] 5) Use LASSO regression algorithm, random forest and SVM-RFE algorithm to screen differentially expressed genes and obtain potential common biomarkers;

[0015] 6) Using the R software package pROC, a nomogram was constructed based on the common biomarkers in step 5), the diagnostic ability of the candidate common biomarker genes was evaluated, and the ROC curve was drawn.

[0016] Furthermore, in the nomogram construction method, the data set in step 1) is preprocessed, including batch effect correction and missing value processing.

[0017] Furthermore, in the nomogram construction method, the batch effect correction uses the empirical Bayes method.

[0018] Furthermore, in the nomogram construction method, the missing value processing includes removing genes and samples with missing values ​​exceeding 50%, and then filling the missing values ​​using the "impute.knn" function in the R package "impute".

[0019] Furthermore, the number of neighbors is set to 10 in the “impute.knn” function.

[0020] Furthermore, in the nomogram construction method, LASSO regression used three-fold cross-validation, random forest used ten-fold cross-validation, and SVM-RFE used five-fold cross-validation to evaluate the diagnostic value of each algorithm.

[0021] Furthermore, in the nomogram construction method, the λ value of LASSO regression was 0.0639847346226388.

[0022] 4. A nomogram for the prediction or diagnosis of osteoporosis associated with atherosclerosis, wherein seven rows of scales are arranged in order from top to bottom:

[0023] The first row of scales is the score scale, with a score range of 0 to 100;

[0024] The second row of scales is the COL1A1 gene expression level, with a value range of 6.5 to 12; 6.5 corresponds to a score of 0 on the scale, and 12 corresponds to a score of 100 on the scale;

[0025] The third row of scales is the SMAD2 gene expression level, with a value range of 5.5 to 4.4; 5.5 corresponds to a score of 0 on the scale, and 4.4 corresponds to a score range of 72-74 on the scale;

[0026] The fourth row of scales is the PTEN gene expression level, with a value range of 5.6 to 7.4; 7.4 corresponds to a score of 0 on the scale, and 5.6 corresponds to a score range of 75-77 on the scale;

[0027] The fifth row of scales is the FBN1 gene expression level, with a value range of 5.5 to 6.9; 6.9 corresponds to a score of 0 on the scale, and 5.5 corresponds to a score range of 37 to 40 on the scale;

[0028] The sixth row of scales is the total score scale, with a value range of 0 to 260; the total score is obtained by summing up the values ​​of the corresponding score scales from the second row to the fifth row;

[0029] The seventh row of scales is the predicted probability of osteoporosis accompanied by atherosclerosis, which is calculated from the total score in the sixth row through the nomogram prediction model.

[0030] Furthermore, the predicted probability of osteoporosis accompanied by atherosclerosis is 0.1 to 0.99, corresponding to a total score scale of 162 to 182.

[0031] 5. A system for predicting or diagnosing osteoporosis accompanied by atherosclerosis, comprising the above-mentioned nomogram.

[0032] The beneficial effects of the present invention are as follows: the present invention provides a nomogram construction method for osteoporosis accompanied by atherosclerosis, and the gene markers for predicting or diagnosing osteoporosis accompanied by atherosclerosis based on the method are COL1A1, SMAD2, PTEN or / and FBN1, and the corresponding nomograms, and a prediction system for the probability of osteoporosis accompanied by atherosclerosis is constructed based on the nomogram, providing a visual basis for the clinical diagnosis of osteoporosis accompanied by atherosclerosis. The present invention adjusts the lambda value in Lasso regression, optimizes the parameter setting of random forest, and the feature selection strategy of SVM-RFE, and finally selects the optimal parameter combination to minimize the prediction error and improve the diagnostic accuracy of the model, and screens out key biomarkers with extremely high diagnostic value. It can be seen from the ROC curve that the common diagnostic ability of the screened biomarkers reaches 0.99, which has extremely high clinical diagnostic application value. Further exploration of the association between osteoporosis and atherosclerosis has important clinical significance for improving the comprehensive management and treatment strategies of elderly patients. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] In order to make the purpose, technical solution and beneficial effects of the present invention clearer, the present invention provides the following drawings for illustration:

[0034] Figure 1 This is a schematic diagram of the research process of the present invention.

[0035] Figure 2 Volcano plot and heat map of differentially expressed genes (DEGs) for the atherosclerosis dataset.

[0036] Figure 3 Volcano plot and heat map of differentially expressed genes (DEGs) in the osteoporosis dataset.

[0037] Figure 4 Analysis of differentially expressed genes (DEGs) by WGCNA.

[0038] Figure 5 Cluster dendrogram of all DEGs.

[0039] Figure 6 Figure 2 is a correlation diagram between modules and atherosclerosis, where A: heat map of the correlation between modules and atherosclerosis, B: scatter plot of the correlation between gene significance and module membership; C: heat map of the correlation between modules and phenotypes.

[0040] Figure 7 Venn diagram for functional enrichment analysis.

[0041] Figure 8 Enrichment analysis of gene intersection in atherosclerosis.

[0042] Fig. 9Venn diagram of differentially expressed genes for osteoporosis and genes for atherosclerosis module.

[0043] Fig.10 Enrichment analysis of common genes in atherosclerosis and osteoporosis.

[0044] Fig.11 This is a protein-protein interaction (PPI) network diagram.

[0045] Fig.12 Visualize the most salient module graph for MCODE plugins.

[0046] Fig.13 are the gene nodes of the first 27 genes in the PPI network.

[0047] Fig.14 For biomarker screening in the Lasso model.

[0048] Fig.15 Biomarker screening for random forest algorithm.

[0049] Fig.16 Biomarker screening for SVM-RFE algorithm.

[0050] Fig.17 Venn diagram of candidate diagnostic genes identified by the three algorithms)

[0051] Fig.18 A visual nomogram for diagnosing osteoporosis with atherosclerosis.

[0052] Fig.19 ROC curves for each candidate gene (COL1A1, SMAD2, PTEN, and FBN1) and the overall results. DETAILED DESCRIPTION

[0053] In order to facilitate the understanding of the present invention, the present invention will be described more comprehensively below in conjunction with the accompanying drawings in the embodiments of the present invention. Preferred embodiments of the present invention are provided in the accompanying drawings. However, the present invention can be implemented in many different forms and is not limited to the embodiments described herein. On the contrary, the purpose of providing these embodiments is to make the understanding of the disclosure of the present invention more thorough and comprehensive.

[0054] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those generally understood by those skilled in the art of the present invention. The terms used herein in the specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. The term "and / or" used herein includes any and all combinations of one or more related listed items. The experimental methods for which specific conditions are not indicated in the embodiments are usually based on conventional conditions or the conditions recommended by the manufacturer.

[0055] The words “include,” “including,” “have,” “contain,” etc. used in this document are open-ended terms, meaning including but not limited to.

[0056] Example 1

[0057] Materials and Methods

[0058] 1.1 Data Collection

[0059] The datasets GSE1428, GSE230665, and GSE100927 were obtained from the GEO database (https: / / www.ncbi.nlm.nih.gov / geo / ) and analyzed based on different microarray platforms. The GSE1428 dataset used the Affymetrix Human Genome U133A Array (GPL96) platform, the GSE230665 dataset used the Agilent-026652 Whole Human Genome Microarray 4×44K v2 (GPL10332) platform, and the GSE100927 dataset was based on the Agilent-039494 SurePrint G3 Human GE v2 8x60K Microarray (GPL17077) platform. The GSE100927 dataset shows transcriptional responses associated with atherosclerosis; GSE230665

[0060] (https: / / www.ncbi.nlm.nih.gov / geo / query / acc.cgi?acc=GSE230665) and GSE1428 datasets focus on the molecular mechanism research of osteoporosis (https: / / www.ncbi.nlm.nih.gov / geo / query / acc.cgi), Figure 1 This is a schematic diagram of the present invention.

[0061] 1.2 Identification of differentially expressed genes (DEGs) between atherosclerosis and osteoporosis

[0062] 1.2.1 Data preprocessing:

[0063] Batch effect correction: The batch effect correction was performed on each dataset using the “SVA” and “Limma” packages in R. The datasets were merged to create a comprehensive data cohort and the Empirical Bayes method was used to eliminate the batch effect.

[0064] Missing value processing: The expression matrix was extracted from the combined osteoporosis dataset and atherosclerosis dataset, and genes and samples with more than 50% missing values ​​were removed to improve the reliability of the data. Subsequently, the “impute.knn” function in the R package “impute” was used to fill the missing values, and the number of neighbors was set to 10 to complete the data completion.

[0065] Data standardization and transformation: In addition, log2 transformation was performed for each gene expression data to reduce the skewness of the data and meet the requirements of normal distribution. In the case where multiple probes identified the same gene, the average expression level of the gene was calculated.

[0066] 1.2.2 Using the Limma package, |log2 fold change (FC)|>1.5 and P<0.05 were set as the criteria for identifying differentially expressed genes (DEGs). The differences in gene expression were evaluated by linear models, and the differentially expressed genes between osteoporosis and normal tissues, and between atherosclerosis and healthy tissues were screened.

[0067] The Limma method was used to analyze the atherosclerosis dataset and a total of 1564 differentially expressed genes were identified, of which 1019 genes were upregulated and 545 genes were downregulated. The volcano plot and heat map of atherosclerosis DEGs can be found in Figure 2 As shown in A and B. In the combined osteoporosis dataset, 2067 DEGs were identified, of which 2059 genes were upregulated and 9 genes were downregulated. Figure 2 Volcano plot and heat map of differentially expressed genes (DEGs) in integrated atherosclerosis dataset. (A) In the volcano plot, red and green circles indicate up-regulated and down-regulated DEGs, respectively. AS stands for atherosclerosis, and DEGs are differentially expressed genes. (B) In the heat map, each row represents a DEG, and each column corresponds to an atherosclerosis case or control sample. Red and blue represent up-regulated and down-regulated gene expression, respectively.

[0068] Volcano plot and heat map of differentially expressed genes DEGs in osteoporosis dataset, see Figure 3As shown in A and B, in the volcano plot A, red and green circles indicate up-regulated and down-regulated DEGs, respectively. DEGs are differentially expressed genes. In the heat map B, each row represents a DEG, and each column corresponds to an osteoporosis case or control sample. Red and blue represent up-regulated and down-regulated gene expression, respectively.

[0069] 1.3 Weighted gene co-expression network analysis (WGCNA) and module gene screening

[0070] Based on the gene expression profile, the median absolute deviation (MAD) of each gene was calculated, and the 50% genes with the smallest MAD were removed. Then, the goodSamplesGenes function in the R software package WGCNA was used to remove abnormal genes and samples. Subsequently, the scale-free co-expression network was constructed using WGCNA. All gene pairs were analyzed using the Pearson correlation matrix and the average connection method, and then a weighted adjacency matrix was constructed by the power function \(A_{mn}=|C_{mn}|^\beta\) (where \(C_{mn}\) represents the Pearson correlation coefficient between genes m and n, and \(A_{mn}\) represents the adjacency value of genes m and n). The β value was set to 12 for soft thresholding to enhance strong correlations and penalize weak correlations. The adjacency matrix was further converted into a topological overlap matrix (TOM), which is used to measure the network connectivity of genes, defined as the sum of the adjacencies of the gene to all genes, and then the corresponding dissimilarity matrix (1-TOM) was calculated. Based on the dissimilarity measure of TOM, an average-linked hierarchical clustering analysis was performed to cluster genes with similar expression patterns into modules, with the minimum number of genes in a module set to 30 and the clustering sensitivity parameter (β) set to 3. In the module analysis, the dissimilarity between module feature genes was calculated, appropriate cut lines were selected to divide the module dendrogram, and some modules were merged, and modules with inter-module distances less than 0.25 were merged.

[0071] Figure 4 Analysis of differentially expressed genes (DEGs) by WGCNA. Figure 4 A and B estimate the soft thresholds for constructing a scale-free co-expression network, where A is a scale-free topological model fitting diagram. This diagram shows whether the network meets the characteristics of a scale-free network under different soft thresholds (power parameters). The soft threshold is usually compared with the correlation fitting index (R 2 ) is presented. The goal is to find a soft threshold so that R 2 Close to or exceeding 0.8, indicating that the network is close to the scale-free property. β = 3 (scale-free R = 0.85) was selected as a "soft" threshold based on scale-freeness and average connectivity. B: Average connectivity map; based on this, 16 gene co-expression modules were generated, Figure 5The cluster dendrogram of all DEGs shows the clustering of atherosclerosis and control samples.

[0072] Figure 6 Heatmap showing the correlation between modules and atherosclerosis, e.g. Figure 6 Shown in different colors in A. Figure 6 Middle B: Scatter plot of correlation between Gene Significance (GS) and Module Membership (MM) in wgcna analysis; Figure 6 Middle C: Module and phenotype correlation heat map. The analysis results of B and C showed that the blue module genes were significantly correlated with atherosclerosis (correlation coefficient = 0.83, P = 1.6e-27), and the blue module genes were also the most significant module genes. The numbers in the brackets at the top and bottom of the figure represent the correlation coefficient and P value, respectively.

[0073] 1.4 Functional enrichment analysis

[0074] The Gene Ontology (GO) system provides structured and computable information about the functions of genes and gene products. Kyoto Encyclopedia of Genes and Genomes (KEGG) is a widely used database for gene system research. Gene mapping was performed based on gene annotations from the R package org.Hs.eg.db (version 3.1.0) and using the latest KEGG pathway gene annotations from the KEGG REST API (https: / / www.kegg.jp / kegg / rest / keggapi.html) as the background set. Functional enrichment analysis was performed using the R package clusterProfiler (version 3.14.3) to obtain gene set enrichment analysis results. The minimum and maximum settings for the gene set were 5 and 5000 genes, respectively. P values ​​less than 0.05 and false discovery rates (FDR) less than 0.1 were considered statistically significant. The analysis was performed in two rounds: the first round was based on the intersection of differentially expressed genes (DEGs) in atherosclerosis and the most significant module genes; the second round was based on the intersection of DEGs in osteoporosis and the most significant module genes in atherosclerosis, performing GO and KEGG analyses, respectively.

[0075] To evaluate whether the dataset GSE10027 reliably reflects the pathogenesis of atherosclerosis, functional enrichment analysis was further performed based on the intersection of Limma and WGCNA module genes. The intersection of 2237 differentially expressed genes (DEGs) from the blue module and 1328 genes generated 236 common genes ( Figure 7 ).

[0076] Figure 8This is the enrichment analysis of gene intersection in atherosclerosis. KEGG pathway analysis shows different significant pathways and related enriched genes in gene intersection, and colors represent different pathways. (AD) shows the GO analysis of gene intersection, including biological processes, cellular components and molecular functions. The y-axis represents different GO terms, the x-axis represents the ratio of genes enriched in the relative GO terms, the size of the circle represents the number of genes, and the color represents the p-value. KEGG: Kyoto Encyclopedia of Genomes and Genomes; GO: Gene Ontology. KEGG analysis showed that common genes were mainly enriched in "cytokine and cytokine receptor interaction" and "chemokine signaling pathway" ( Figure 8 GO analysis showed that in the cellular component (CC) ontology, these common genes were mainly located in “extracellular” and “vesicle” ( Figure 8 In the biological process (BP) entries, common genes were enriched in "regulation of response to estrogen" and "stress response" ( Figure 8 C). The results of molecular function (MF) analysis showed that "signal receptor binding" and "extracellular matrix structure" were the most significant items in the common genes ( Figure 8 (middle D).

[0077] 1.5 Construction of protein-protein interaction network

[0078] To mine interactions between protein-coding genes, a protein-protein interaction (PPI) network was constructed using the String database (version 11.5, website: www.string-db.org), with a minimum interaction score of 0.400. Subsequently, the images downloaded from String were modified using Cytoscape software, and the MCODE plug-in was used to identify important interacting genes. All genes that were able to interact in the PPI network were selected for subsequent analysis.

[0079] To further explore whether the key genes associated with atherosclerosis are related to the pathogenesis of osteoporosis, we used a Venn diagram to display the intersection of differentially expressed genes (DEGs) of osteoporosis and genes in the atherosclerosis module, and identified 281 common genes ( Fig. 9 ). KEGG enrichment analysis showed that these genes were mainly enriched in "human papillomavirus infection" and "human T-cell leukemia virus 1 infection" ( Fig.10 GO analysis showed that these genes were mainly enriched in "endomembrane system" and "organelle membrane" in cellular components (CC); in biological processes (BP), they were enriched in "response to organic matter" and "positive regulation of biosynthesis"; in molecular functions (MF), these genes were enriched in "RNA polymerase II proximal promoter sequence-specific DNA binding" and "protease binding" ( Fig.10 A: CC, cellular component, B: BP, biological process. C: MF, molecular function. D: KEGG, Kyoto Encyclopedia of Genes and Genomes.

[0080] After confirming the screened genes, a protein-protein interaction (PPI) network was constructed ( Fig.11 ) to identify interacting node genes for subsequent machine learning screening. The top 27 genes were ranked according to the number of nodes ( Fig.13 ), and visualized the most active modules using the MCODE plugin ( Fig.12 ).

[0081] 1.6 Machine Learning

[0082] In order to further screen more accurate candidate genes for diagnosing atherosclerosis and osteoporosis, this study used multiple machine learning algorithms for comprehensive screening.

[0083] First, LASSO regression analysis was used. This method aims to improve prediction accuracy through variable selection. It is also an effective regularization technique that enhances the interpretability of statistical models. Gene expression data were integrated using the R package glmnet, and the LASSO-Cox method was applied for regression analysis. Three-fold cross validation was set to obtain the best model. The data set was evenly and randomly divided into three parts, and the model was evaluated by three-fold cross validation. In the three-fold cross validation, two of the data were selected for model training each time, and the remaining one was used for test validation. The whole process was repeated three times to ensure that each data had a chance to be used as a test set. In each iteration, LASSO-Cox regression was used to train the model according to different regularization parameters λ, and the model performance was evaluated by the prediction error of the test set. After multiple iterations, the final λ value was determined to be 0.0639847346226388, which minimized the cross validation error.

[0084] Secondly, the support vector machine recursive feature elimination (SVM-RFE) method was used, which aims to improve the predictive performance and interpretability of the model by recursively removing unimportant features. In this study, by reading gene expression data and implementing five-fold cross-validation, the SVM-RFE algorithm gradually eliminated unimportant features and extracted the main variables. Subsequently, the first 9 features were selected for model construction, and the impact of different numbers of features on model performance was evaluated, so as to finally identify important genes related to the disease.

[0085] Finally, the recursive feature elimination (RFE) method combined with the random forest algorithm was used for feature selection to screen gene features related to the research objectives. First, the data was read from the CSV file and the random seed was set to ensure the reproducibility of the results. Then, the parameters of RFE were configured using the "rfeControl" function, and the random forest algorithm (rfFuncs) and ten-fold repeated cross validation (LGOCV) were specified for model evaluation. The features were selected through the "rfe" function, the feature data and the target variable (gene classification label) were input, and the range of different feature numbers was set for step-by-step evaluation. Finally, the gene features with the best predictive accuracy were extracted and the results were saved in the output file. By comprehensively applying the above three machine learning methods, extracting their respective results and calculating their intersection, a group of key hub genes confirmed as the diagnosis of atherosclerosis and osteoporosis were finally identified.

[0086] By adjusting the λ value in Lasso regression, the optimal parameter combination is finally selected to minimize the prediction error and improve the diagnostic accuracy of the model, thereby screening out key biomarkers with high diagnostic value.

[0087] As mentioned above, we applied three machine learning algorithms, namely LASSO regression, random forest (RF), and support vector machine-recursive feature elimination (SVM-RFE), to screen candidate genes that can be accurately used for diagnostic value assessment and nomogram construction. Fig.14 As shown in A and B, the LASSO regression algorithm identified 5 potential candidate biomarkers; the RF algorithm ranked each gene according to its importance and identified 8 potential candidate biomarkers ( Fig.15 ). The SVM-RFE algorithm identified seven potential candidate biomarkers ( Fig.16 The intersection of the 8 candidate genes identified by the RF algorithm, the 5 candidates identified by LASSO, and the 7 genes identified by SVM-RFE was visualized through the Venn diagram ( Fig.17 ), and ultimately four genes (COL1A1, SMAD2, PTEN, and FBN1) were identified for final validation.

[0088] Example 2

[0089] Nomogram construction and receiver operating characteristic assessment

[0090] The construction of nomograms has certain value in the diagnosis of clinical atherosclerosis and osteoporosis. Based on candidate genes, the nomogram was constructed using the R package pROC (version 1.17.0.1). "Score" represents the score of the candidate gene, while "Total Score" represents the sum of the scores of all the above genes. Subsequently, receiver operating characteristic (ROC) curves were established using the GSE35958 (osteoporosis) and GSE28829 (atherosclerosis) datasets to evaluate the diagnostic value of candidate genes and nomograms in the diagnosis of atherosclerosis and osteoporosis.

[0091] A nomogram was constructed based on four candidate hub genes, such as Fig.18 As shown, ROC curves were established to evaluate the diagnostic specificity and sensitivity of each gene and nomogram. Fig.19The area under the curve (AUC) and its 95% confidence interval (CI) were calculated for each candidate gene (COL1A1, SMAD2, PTEN, and FBN1), where A: COL1A1; B: SMAD2; C: PTEN.D: FBN1; E: total ROC, the results are as follows: COL1A1 (AUC 0.93, CI 0.98-0.87), SMAD2 (AUC 0.82, CI 0.91-0.74), PTEN (AUC 0.88, CI 0.95-0.81), FBN1 (AUC 0.72, CI 0.82-0.61), four proteins combined ROC (AUC0.99, CI 1.00-0.98). All candidate genes showed high diagnostic value in atherosclerosis associated with osteoporosis, among which the constructed nomogram showed the highest diagnostic value. These results show that these genes have good clinical application potential in identifying atherosclerosis associated with osteoporosis, providing an important basis for further diagnosis and treatment. In some practical detection applications, any suitable technology that allows qualitative and / or quantitative assessment of the level of biomarker genes in the sample to be detected can be used. Comparison can be made by reference to a standard control or a control level found in healthy tissue, and the level of transcribed genes can be determined by Northern blotting and / or RT-PCR. Real-time PCR quantitative detection can also be used, and quantitative analysis of gene expression can be achieved by using appropriate primers for the gene of interest. Immunoassay methods such as immunohistochemistry, immunoblotting, protein blotting, dot blot, enzyme-linked immunosorbent assay (ELISA), radioimmunoassay (RIA), enzyme immunoassay and other assay methods can also be used. Immunohistochemistry and immunocytochemistry refer to the process of locating proteins in tissue sections or cells, respectively, through the principle of antigen binding to their corresponding antibodies in tissues or cells, and further visualization is achieved by labeling antibodies with color-producing or fluorescent labels. Color labels such as horseradish peroxidase and alkaline phosphatase are used. Fluorescent labels such as fluorescein isothiocyanate (FITC) or phycoerythrin (PE) can also be used to detect protein gene expression products in some embodiments, such as matrix-assisted laser desorption / ionization time of flight (MALDI-TOF MS) and electrospray mass spectrometry (ESMS).

[0092] The ROC curve and nomogram of each candidate gene (COL1A1, SMAD2, PTEN, and FBN1) showed significant diagnostic value for osteoporosis with atherosclerosis. COL1A1: collagen type I alpha 1 chain; SMAD2, maternal adult stem cell-associated protein 2; PTEN: phosphatase and tensin homolog; FBN1: fibrinogen 1. The nomogram visually displays the contribution of each gene in disease diagnosis. The "total score" is calculated, which is the sum of the scores of all candidate genes. Provides a visual basis for clinical decision-making.

[0093] A system for predicting or diagnosing osteoporosis with atherosclerosis can be established, by corresponding the expression values ​​of the four genes COL1A1, SMAD2, PTEN, and FBN1 to corresponding scores, and then adding up the scores to finally obtain the probability of osteoporosis with atherosclerosis.

[0094] The system includes a nomogram with seven rows of scales arranged from top to bottom:

[0095] The first row of scales is the score scale, with a score range of 0 to 100;

[0096] The second row of scales is the COL1A1 gene expression level, with a value range of 6.5 to 12; 6.5 corresponds to a score of 0 on the scale, and 12 corresponds to a score of 100 on the scale;

[0097] The third row of scales is the SMAD2 gene expression level, with a value range of 5.5 to 4.4; 5.5 corresponds to a score of 0 on the scale, and 4.4 corresponds to a score range of 72-74 on the scale;

[0098] The fourth row of scales is the PTEN gene expression level, with a value range of 5.6 to 7.4; 7.4 corresponds to a score of 0 on the scale, and 5.6 corresponds to a score range of 75-77 on the scale;

[0099] The fifth row of scales is the FBN1 gene expression level, with a value range of 5.5 to 6.9; 6.9 corresponds to a score of 0 on the scale, and 5.5 corresponds to a score range of 37 to 40 on the scale;

[0100] The sixth row of scales is the total score scale, with a value range of 0 to 260; the total score is obtained by summing up the values ​​of the corresponding score scales from the second row to the fifth row;

[0101] The seventh row scale is the predicted probability of osteoporosis with atherosclerosis, which is calculated by the sixth row total score through the nomogram prediction model. The predicted probability of osteoporosis with atherosclerosis is 0.1 to 0.99, corresponding to 162 to 182 of the total score scale.

[0102] Finally, it should be noted that the above preferred embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail through the above preferred embodiments, for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as the scope of protection defined by the claims.

Claims

1. A gene marker for predicting or diagnosing osteoporosis associated with atherosclerosis, characterized in that: The gene markers are COL1A1, SMAD2, PTEN and / or FBN1.

2. Use of a reagent for detecting the expression level of the gene marker according to claim 1 in the preparation of a product for predicting or diagnosing osteoporosis associated with atherosclerosis.

3. A method for constructing a nomogram for osteoporosis associated with atherosclerosis, characterized in that: The following steps are involved: 1) Obtain GSE1428, GSE230665 and GSE1000927 datasets from the GEO database; 2) Using the Limma package, we set |log2 fold change (FC)|>1.5 and P<0.05 to screen the differentially expressed genes between osteoporosis and normal tissues, and between atherosclerosis and healthy tissues; 3) Identify candidate gene modules for atherosclerosis using weighted gene co-expression network analysis based on Pearson correlation; 4) Use the STRING database to construct a protein-protein interaction network and perform functional enrichment analysis; 5) Use LASSO regression algorithm, random forest and SVM-RFE algorithm to screen differentially expressed genes and obtain potential common biomarkers; 6) Using the R software package pROC, a nomogram was constructed based on the common biomarkers in step 5), the diagnostic ability of the candidate common biomarker genes was evaluated, and the ROC curve was drawn.

4. The method for constructing a nomogram according to claim 3, characterized in that: In step 1), the data set is preprocessed, including batch effect correction and missing value processing.

5. The method for constructing a nomogram according to claim 4, characterized in that: The missing value processing includes removing genes and samples with missing values ​​exceeding 50%, and then filling the missing values ​​using the "impute.knn" function in the R package "impute".

6. The method for constructing a nomogram according to claim 3, characterized in that: Three-fold cross validation was used for LASSO regression, ten-fold cross validation was used for random forest, and five-fold cross validation was used for SVM-RFE to evaluate the diagnostic value of each algorithm.

7. The method for constructing a nomogram according to claim 3, characterized in that: The lambda value for LASSO regression is 0.0639847346226388.

8. A nomogram obtained by the construction method according to any one of claims 2 to 7, characterized in that: The nomogram has seven rows of scales arranged from top to bottom: The first row of scales is the score scale, with a score range of 0 to 100; The second row of scales is the COL1A1 gene expression level, with a value range of 6.5 to 12; 6.5 corresponds to a score of 0 on the scale, and 12 corresponds to a score of 100 on the scale; The third row of scales is the SMAD2 gene expression level, with a value range of 5.5 to 4.4; 5.5 corresponds to a score of 0 on the score scale, and 4.4 corresponds to a score range of 72-74 on the score scale; The fourth row of scales is the PTEN gene expression level, with a value range of 5.6 to 7.4; 7.4 corresponds to a score of 0 on the score scale, and 5.6 corresponds to a score range of 75-77 on the score scale; The fifth row of scales is the FBN1 gene expression level, ranging from 5.5 to 6.9; 6.9 corresponds to a score of 0 on the score scale, and 5.5 corresponds to a score range of 37 to 40 on the score scale; The sixth row of scales is the total score scale, with a value range of 0 to 260; the total score is obtained by summing up the values ​​of the corresponding score scales from the second row to the fifth row; The seventh row of scales is the predicted probability of osteoporosis accompanied by atherosclerosis, which is calculated from the total score in the sixth row through the nomogram prediction model.

9. The system for predicting or diagnosing osteoporosis accompanied with atherosclerosis according to claim 8, characterized in that: The predicted probability of osteoporosis accompanied by atherosclerosis is 0.1 to 0.99, corresponding to a total score of 162 to 182.

10. A system for predicting or diagnosing osteoporosis associated with atherosclerosis, characterized in that: Comprising the nomogram described in claim 8 or 9.

Images