Gastric cancer biomarker mining method and system

Abstract

The invention belongs to the technical field of biological medicine, and relates to a gastric cancer biomarker mining method and system. The method comprises the following steps: acquiring data; preprocessing the data; generating a gene interaction network; performing data difference analysis; performing function enrichment analysis; analyzing a survival prognosis model gene; carrying out GSEA enrichment analysis; carrying out progression-free lifetime analysis; a Nomogram model is constructed; analyzing the causal relationship between the gastric cancer biomarker and the gastric cancer; analyzing the causal relationship between protease and gastric cancer; and analyzing the causal relationship between the gastric cancer biomarker and the protease. According to the invention, the gastric cancer mechanism is explored by combining a bioinformatics method and a Mendel randomization method, and the gastric cancer biomarker indirectly affects the occurrence of gastric cancer through mediated protease, so that the mining of the gastric cancer biomarker is realized.

Description

Technical Field

[0001] The present invention belongs to the field of biomedicine technology, and specifically, relates to a method and system for mining gastric cancer biomarkers. Background Art

[0002] Currently, metabolomics methods are used to mine gastric cancer biomarkers, and machine learning algorithms are used to screen out six specific gastric cancer biomarkers. This method only mines gastric cancer biomarkers through a single mining method, which has certain limitations, and there is no relevant evidence for the predictive and diagnostic effect of BRCA1 on gastric cancer. Summary of the invention

[0003] In order to solve the above technical problems, the present invention provides a method and system for mining gastric cancer biomarkers.

[0004] In a first aspect, the present invention provides a method for mining gastric cancer biomarkers, comprising:

[0005] Data acquisition: Obtain gene expression data from the TCGA database, download gastric cancer clinical data, and download data sets from the GEO database;

[0006] Data preprocessing: By setting the differential expression threshold, differentially expressed genes are determined to obtain integrated gene expression data;

[0007] Gene interaction network generation: Based on the integrated gene expression data, by integrating the interaction network between genes, the topological parameters of each node in the interaction network are calculated, the topological parameters are sorted by size, and the core gene target network is determined based on the set threshold.

[0008] Data difference analysis: Based on the data set in the GEO database, the core gene target network analysis was performed, and the differential expression of the core genes in the intergroup relationship was analyzed by identifying the differential expression of the core genes in gastric cancer tissues and normal tissues, and the statistical differences were evaluated by calculating the P value;

[0009] Functional enrichment analysis: Based on the statistical difference evaluation results, the core gene target network was analyzed for overall GO and KEGG enrichment, and the relationship between the core gene target network and individual core gene targets was analyzed by evaluating the GO and KEGG enrichment of individual core genes;

[0010] Gene analysis of survival prognosis model: Based on TCGA clinical survival data, the Kaplan-Meier method was used to compare the survival of core genes in normal tissues and gastric cancer tissues. The statistical difference and functional enrichment analysis results were combined to determine the candidate key genes that affect the occurrence and survival of gastric cancer, and to construct a survival prognosis model.

[0011] GSEA enrichment analysis: GSEA enrichment analysis was performed on the survival prognosis model, and GSEA-GO and GSEA-KEGG results were calculated;

[0012] Progression-free survival analysis: Based on TCGA clinical data, the survival prognostic model was analyzed for progression-free survival, and the specific parameters, sensitivity parameters, and specificity parameters of the survival prognostic model were evaluated;

[0013] Nomogram model construction: Based on the gastric cancer dataset of the TCGA database, a Nomogram model was constructed for gastric cancer biomarkers to evaluate the clinical relevance of the survival prognostic model in disease stage, age, gender, grade and TMN stage, and calculate the survival rate;

[0014] Analysis of the causal relationship between gastric cancer biomarkers and gastric cancer: Based on the GWAS Catalog database and the IEUOpenGWAS database, the causal relationship between gastric cancer biomarkers and gastric cancer was calculated by Mendelian randomization method, with gastric cancer biomarkers as exposure factors and gastric cancer as outcome factors, and exchange verification was performed;

[0015] Analysis of the causal relationship between protease and gastric cancer: Using the Mendelian randomization method, protease was used as the exposure factor and gastric cancer was used as the outcome factor to calculate the causal relationship between protease and gastric cancer, and exchange verification was performed;

[0016] Analysis of the causal relationship between gastric cancer biomarkers and proteases: Using the Mendelian randomization method, with gastric cancer biomarkers as exposure factors and proteases as outcome factors, the causal relationship between gastric cancer biomarkers and proteases was calculated and exchange verification was performed.

[0017] In a second aspect, the present invention provides a gastric cancer biomarker mining system, comprising a data acquisition unit, a data preprocessing unit, a gene interaction network generation unit, a data difference analysis unit, a function enrichment analysis unit, a survival prognosis model gene analysis unit, a GSEA enrichment analysis unit, a progression-free survival analysis unit, a Nomogram model construction unit, a first causal relationship analysis unit, a second causal relationship analysis unit, and a third causal relationship analysis unit;

[0018] A data acquisition unit is used to acquire gene expression data from the TCGA database, download clinical data of gastric cancer, and download data sets from the GEO database;

[0019] A data preprocessing unit, used to determine differentially expressed genes by setting differential expression thresholds, and obtain integrated gene expression data;

[0020] The gene interaction network generation unit is used to calculate the topological parameters of each node in the interaction network based on the integrated gene expression data by integrating the interaction network between genes, sort the topological parameters by size, screen them according to the set threshold, and determine the core gene target network;

[0021] Data difference analysis unit, used to perform core gene target network analysis based on the data set in the GEO database, analyze the differential expression of core genes in the intergroup relationship by identifying the differential expression of core genes in gastric cancer tissues and normal tissues, and evaluate the statistical difference by calculating the P value;

[0022] Functional enrichment analysis unit, used to perform overall GO and KEGG enrichment analysis on the core gene target network based on the statistical difference evaluation results, and to analyze the relationship between the core gene target network and individual core gene targets by evaluating the GO and KEGG enrichment of individual core genes;

[0023] The survival prognosis model gene analysis unit is used to compare the survival of core genes in normal tissues and gastric cancer tissues based on TCGA clinical survival data using the Kaplan-Meier method, identify candidate key genes that affect the occurrence and survival of gastric cancer, and construct a survival prognosis model;

[0024] GSEA enrichment analysis unit, used to perform GSEA enrichment analysis on the survival prognosis model and calculate GSEA-GO and GSEA-KEGG results;

[0025] Progression-free survival analysis unit, used to perform progression-free survival analysis on the survival prognostic model based on TCGA clinical data, and to evaluate the specific parameters, sensitivity parameters, and specificity parameters of the survival prognostic model;

[0026] The Nomogram model building unit is used to build a Nomogram model for gastric cancer biomarkers based on the gastric cancer dataset of the TCGA database, evaluate the clinical relevance of the survival prognostic model in disease stage, age, gender, grade and TMN stage, and calculate the survival rate;

[0027] The first causal relationship analysis unit is used to calculate the causal relationship between gastric cancer biomarkers and gastric cancer based on the GWAS Catalog database and the IEU OpenGWAS database through the Mendelian randomization method, taking gastric cancer biomarkers as exposure factors and gastric cancer as outcome factors, and perform exchange verification;

[0028] The second causal relationship analysis unit was used to calculate the causal relationship between protease and gastric cancer by using the Mendelian randomization method, taking protease as the exposure factor and gastric cancer as the outcome factor, and to perform exchange verification;

[0029] The third causal relationship analysis unit is used to apply the Mendelian randomization method, taking the gastric cancer biomarker as the exposure factor and the protease as the outcome factor, calculate the causal relationship between the gastric cancer biomarker and the protease, and perform exchange verification.

[0030] Based on the above technical solution, the present invention can also be improved as follows.

[0031] Furthermore, the integrated gene expression data were imported into the String database to calculate the relationships, construct the PPI network to obtain the interaction network, and use the CytoHubba plug-in in the Cytoscape software to screen the core gene target network to obtain several core genes.

[0032] Furthermore, during the functional enrichment analysis, the core gene target network was subjected to overall GO and KEGG enrichment analysis according to biological process, cellular component, molecular function and pathway.

[0033] Furthermore, a survival prognosis model was constructed and the risk score was calculated to determine the accuracy of the survival prognosis model, including: assuming that the risk score is risk score, i is the sample number, n is the total number of sample numbers, coefi is the score coefficient obtained by survival analysis, and id is the expression level of the gastric cancer biomarker gene, then:

[0034]

[0035] Furthermore, five Mendelian algorithms, including MR Egger algorithm, weighted median algorithm, IVW algorithm, simple mode algorithm and weighted mode algorithm, were used to calculate the causal relationship between gastric cancer biomarkers and gastric cancer, the causal relationship between proteases and gastric cancer, and the causal relationship between gastric cancer biomarkers and proteases, respectively.

[0036] The beneficial effect of the present invention is that the present invention uses bioinformatics methods in combination with Mendelian randomization methods to explore the mechanism of gastric cancer and realizes the mining of gastric cancer biomarkers. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 A schematic diagram of a gastric cancer biomarker mining method provided in Example 1 of the present invention;

[0038] Figure 2 This is a schematic diagram of a gastric cancer biomarker mining system provided in Example 2 of the present invention. DETAILED DESCRIPTION

[0039] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings here can be arranged and designed in various different configurations.

[0040] Example 1

[0041] Based on the same principle as the method shown in Example 1 of the present invention, an embodiment of the present invention further provides a method for mining gastric cancer biomarkers, comprising:

[0042] Data acquisition: Obtain gene expression data from the TCGA (The Cancer Genome Atlas) database, download gastric cancer clinical data, and download data sets from the GEO database (Gene Expression Omnibus);

[0043] Data preprocessing: By setting the differential expression threshold, differentially expressed genes are determined to obtain integrated gene expression data;

[0044] Gene interaction network generation: Based on the integrated gene expression data, by integrating the interaction network between genes, the topological parameters of each node in the interaction network are calculated, the topological parameters are sorted by size, and the core gene target network is determined based on the set threshold.

[0045] Data difference analysis: Based on the data set in the GEO database, the core gene target network analysis was performed, and the differential expression of the core genes in the intergroup relationship was analyzed by identifying the differential expression of the core genes in gastric cancer tissues and normal tissues, and the statistical differences were evaluated by calculating the P value;

[0046] Functional enrichment analysis: Based on the statistical difference evaluation results, the core gene target network was subjected to overall GO (Gene Ontology) and KEGG (Kyoto Encyclopedia of Genes and Genomes) enrichment analysis, and the relationship between the core gene target network and individual core gene targets was analyzed by evaluating the GO and KEGG enrichment of individual core genes;

[0047] Gene analysis of survival prognosis model: Based on TCGA clinical survival data, the Kaplan-Meier method was used to compare the survival of core genes in normal tissues and gastric cancer tissues. The statistical difference and functional enrichment analysis results were combined to determine the candidate key genes that affect the occurrence and survival of gastric cancer, and to construct a survival prognosis model.

[0048] GSEA (Gene Set Enrichment Analysis) enrichment analysis: GSEA enrichment analysis was performed on the survival prognosis model, and GSEA-GO and GSEA-KEGG results were calculated;

[0049] Progression-free survival analysis: Based on TCGA clinical data, the survival prognostic model was analyzed for progression-free survival, and the specific parameters, sensitivity parameters, and specificity parameters of the survival prognostic model were evaluated;

[0050] Nomogram model construction: Based on the gastric cancer dataset of the TCGA database, a Nomogram model was constructed for gastric cancer biomarkers to evaluate the clinical relevance of the survival prognostic model in disease stage, age, gender, grade and TMN stage, and calculate the survival rate;

[0051] Analysis of the causal relationship between gastric cancer biomarkers and gastric cancer: Based on the GWAS Catalog database and the IEUOpenGWAS database, the causal relationship between gastric cancer biomarkers and gastric cancer was calculated by Mendelian randomization method, with gastric cancer biomarkers as exposure factors and gastric cancer as outcome factors, and exchange verification was performed;

[0052] Analysis of the causal relationship between protease (reduced coenzyme II) and gastric cancer: Using the Mendelian randomization method, protease was used as the exposure factor and gastric cancer was used as the outcome factor to calculate the causal relationship between protease and gastric cancer, and exchange verification was performed;

[0053] Analysis of the causal relationship between gastric cancer biomarkers and proteases: Using the Mendelian randomization method, taking gastric cancer biomarkers as exposure factors and proteases as outcome factors, the causal relationship between gastric cancer biomarkers and proteases was calculated and exchanged for verification;

[0054] According to the causal relationship between gastric cancer biomarkers and gastric cancer, the causal relationship between proteases and gastric cancer, and the causal relationship between gastric cancer biomarkers and proteases, it was found that gastric cancer biomarkers mediate proteases, destroy cellular oxidative stress, and destroy cellular balance, thereby leading to the occurrence of gastric cancer. Reverse Mendelian randomization analysis showed that gastric cancer biomarkers and gastric cancer have no causal relationship, that is, gastric cancer biomarkers indirectly affect the occurrence of gastric cancer by mediating proteases, rather than directly intervening through gastric cancer biomarkers to cause the occurrence of gastric cancer.

[0055] Taking BRCA1 as an example of gastric cancer biomarkers, BRCA1 is an important tumor suppressor gene, which consists of 1863 residues of nuclear polymorphism, with a prominent n-terminal RING domain (the presence of a RING finger-shaped domain is a RING-type E3 ubiquitin protein ligase, which transports ubiquitin from the E2 enzyme to the substrate protein) and an evolutionarily conserved c-terminal BRCT (BRCA1 C-terminal, BRCA1C-terminal) domain. It is worth noting that 20%-30% of familial breast cancer and ovarian cancer are related to BRCA1 mutations. There is evidence that the BRCA1 gene is not only related to breast cancer and ovarian cancer; it also shows significant differences in various other types of cancer. Based on the above theory, the present invention integrates databases to identify the expression differences of BRCA1 in gastric cell carcinoma tumor tissues and normal tissues and further verify its mechanism of action. The present invention found that by analyzing clinical samples and constructing a prognostic model, it was found that low expression of BRCA1 would lead to a high risk of gastric cancer, and according to survival analysis, long-term low expression of BRCA1 would lead to a much lower survival time of gastric cancer patients than the normal group. In addition, through Mendelian randomization analysis, BRCA1 causes gastric cancer by mediating NADPH oxidoreductase, destroying the redox homeostasis in the body, thereby leading to the occurrence of gastric cancer.

[0056] The present invention combines gastric cell specimens from gastric cancer patients and healthy control patients, screens genes that may cause gastric cancer based on bioinformatics methods as gastric cancer biomarkers, and then explores factors that cause gastric cancer based on a two-sample mediated Mendelian randomization method.

[0057] There is a significant difference in the expression of BRCA1 in gastric cancer tissues compared with normal gastric tissues and gastric cancer tissues. Gene expression correlation analysis showed that BRCA1 was highly expressed in gastric cancer tissues and lowly expressed in normal tissues. According to survival analysis and clinical sample analysis, the survival rate of gastric cancer in the BRCA1 high expression group was much higher than that in the low expression group, and the prevalence rate in the group under 65 years old was much higher than that in the group over 65 years old. Through Mendelian randomization analysis, it was found that there was no direct causal relationship between BRCA1 and GC, but BRCA1 had a strong causal relationship with NADPH, and NADPH also had a strong causal relationship with GC. Reverse Mendelian randomization also strengthened this conjecture. The results showed that BRCA1 indirectly regulates the expression of NADPH, leading to the occurrence of gastric cancer.

[0058] The present invention integrates four different types of databases, namely the GEO database, the TCGA database, the GWAS Catalog database and the IEU OpenGWAS database (Integrative Epidemiology Unit (IEU) openGWAS project, an open genome-wide association study data set) to identify the expression difference of BRCA1 in gastric cell carcinoma tumor tissues and normal tissues and further verify its mechanism of action.

[0059] During the data preprocessing process, “gastric cancer second-generation sequencing dataset TCGA-STAD” was used as the keyword to obtain gene expression data from the TCGA database, and gastric cancer clinical data were downloaded. At the same time, in order to make the dataset more diverse, datasets GSE13911 and GSE29272 were downloaded from the GEO database, and differentially expressed genes were found by setting differential expression domain values.

[0060] Optionally, the integrated gene expression data are imported into the String database to calculate the mutual relationships, construct a PPI network to obtain the interaction network, and use the CytoHubba plug-in in the Cytoscape software to screen the core gene target network to obtain several core genes.

[0061] In the process of data difference analysis, the core gene target network was further analyzed based on the data sets GSE13911 and GSE29272 in the GEO database. The differential expression of core genes in the inter-group relationship was determined by identifying the differential expression of core genes in gastric cancer tissues and normal tissues, and the statistical difference was evaluated by calculating the P value. If the P value is less than 0.05, there is a statistical difference, or p<0.01 (there is a significant statistical difference), or p<0.001 (there is an extremely significant statistical difference). By evaluating the difference in expression of the screened core genes in the low-risk group and the high-risk group, either high expression in the high-risk group or high expression in the low-risk group, the high and low expression of the core genes in normal tissues and gastric cancer tissues are intuitively displayed.

[0062] Optionally, during the functional enrichment analysis, the core gene target network is subjected to overall GO and KEGG enrichment analysis according to biological process, cellular component, molecular function and pathway.

[0063] ClusterProfiler (an efficient cluster analysis tool in bioinformatics) was used to perform GO and KEGG enrichment analysis on the core gene network as a whole, and the relationship between the core gene network and individual core genes was analyzed by evaluating the GO and KEGG enrichment of individual core genes. The results of the enrichment analysis can be used to determine the functions of the core network target genes and the pathways that may act on the development of gastric cancer.

[0064] Based on TCGA clinical survival data, the core genes were further analyzed. The Kaplan-Meier method was used to compare the survival of the core genes in normal tissues and gastric cancer tissues, identify whether the core genes affect the occurrence and survival of gastric cancer, find candidate key genes, and construct a prognostic model.

[0065] Optionally, a survival prognosis model is constructed, and a risk score is calculated to determine the accuracy of the survival prognosis model, including: assuming that the risk score is risk score, i is the sample number, n is the total number of sample numbers, coefi is the score coefficient obtained by survival analysis, and id is the expression level of the gastric cancer biomarker gene, then:

[0066]

[0067] Based on the prognostic model, GSEA enrichment analysis was performed on the prognostic model through the "org.Hs.Eg.Db" database file, and the GSEA-GO and GSEA-KEGG results were calculated.

[0068] The survival prognostic model was evaluated by GSEA-GO and GSEA-KEGG, and it was found that BRCA1 affects the development of gastric cancer mainly on the cell cycle, olfactory transduction and ubiquitin-mediated proteolysis.

[0069] The survival model was analyzed for progression-free survival based on TCGA clinical data, and the specificity and sensitivity of the survival model prognosis were evaluated using the receiver operating characteristic curve. The area under the receiver operating characteristic curve was evaluated to determine whether it was considered robust. The timeRoc package was used to calculate the 1-year, 3-year, and 5-year prognoses.

[0070] Based on the gastric cancer dataset of TCGA database, the clinical relevance of the prognostic model in disease stage, age, gender, grade and TMN stage was evaluated, and the survival rate was calculated.

[0071] Optionally, five Mendelian algorithms are used, including MR Egger algorithm, weighted median algorithm, IVW algorithm (the random-effects inverse variance-weighted algorithm, referred to as IVW algorithm), simple mode algorithm and weighted mode algorithm, to calculate the causal relationship between gastric cancer biomarkers and gastric cancer, the causal relationship between proteases and gastric cancer, and the causal relationship between gastric cancer biomarkers and proteases, respectively.

[0072] Analysis of the causal relationship between gastric cancer biomarkers and gastric cancer: Based on the databases The NHGRI-EBI Catalog of human genome-wide association studies (GWAS Catalog) and the IEU OpenGWAS, the gastric cancer biomarkers were used as exposure factors and gastric cancer was used as outcome factors through Mendelian randomization. The causal relationship was calculated using the MR Egger algorithm, weighted median, IVW algorithm, simple mode algorithm, and weighted mode algorithm, respectively. The random-effects inverse variance-weighted algorithm (IVW algorithm) and MR Egger algorithm were used to analyze heterogeneity, sensitivity, and specificity. At the same time, in order to verify whether gastric cancer as an exposure factor interferes with gastric cancer biomarkers as outcome factors, the two were exchanged for verification.

[0073] Analysis of the causal relationship between protease and gastric cancer: Using the Mendelian randomization method, protease was used as an exposure factor and gastric cancer was used as an outcome factor. The causal relationship was calculated using the above five Mendelian algorithms, and the IVW and MR Egger algorithms were used to analyze heterogeneity, sensitivity, and specificity. At the same time, in order to verify whether gastric cancer as an exposure factor interfered with protease as an outcome factor, the two were exchanged for verification.

[0074] Analysis of the causal relationship between gastric cancer biomarkers and proteases: Using the Mendelian randomization method, gastric cancer biomarkers were used as exposure factors and proteases were used as outcome factors. The above five Mendelian algorithms were used to calculate the causal relationship, and IVW and MR Egger algorithms were used to analyze heterogeneity, sensitivity, and specificity. At the same time, in order to verify whether proteases as exposure factors interfere with gastric cancer biomarkers as outcome factors, the two were exchanged for verification.

[0075] The experimental process of the present invention is as follows:

[0076] (1) Download the gene expression data of gastric cancer from the TCGA database, and download the GSE13911 and GSE29272 datasets through GEO. After setting the difference threshold, 2536 differentially expressed genes were obtained in GSE13911, 6208 differentially expressed genes were obtained in GSE29272, and 11258 differentially expressed genes were obtained in TCGA. The datasets were intersected to obtain a total of 792 intersection genes;

[0077] (2) The integrated gene expression data were imported into the String database to calculate the relationships, construct the PPI network, and use the CytoHubba plug-in in the Cytoscape software to screen the core gene target network, resulting in a total of 9 core genes;

[0078] (3) The two datasets in the GEO database were merged and de-batched to form a new dataset. The core network genes were further analyzed based on the new dataset to analyze the differential expression of the core genes in gastric cancer tissues and normal tissues;

[0079] (4) Conduct GO enrichment analysis and KEGG enrichment analysis on the core network genes, and explore the mechanism by which the core genes affect the development of gastric cancer by analyzing biological processes, cellular components, molecular functions, and pathways;

[0080] (5) Construct a prognostic model to analyze the prognosis of core genes, and calculate the risk score to determine the accuracy of the gene prognostic model. Assume that the risk score is risk score, i is, coefi is, id is, and the risk score is calculated as follows:

[0081]

[0082] Combining the differential expression between the normal group and the gastric cancer group and the risk score, the Kaplan-Meier survival analysis curve was constructed, and it was concluded that BRCA1 has great value in prognosis, and subsequent analysis was performed;

[0083] (6) GSEA enrichment analysis was performed on the BRCA1 gene, and the prognostic model was evaluated by GSEA-GO and GSEA-KEGG, showing that BRCA1 affects the development of gastric cancer mainly in the cell cycle, olfactory transduction, and ubiquitin-mediated proteolysis;

[0084] (7) Analyze the progression-free survival of BRCA1 and calculate the 1-year, 3-year, and 5-year prognostic results, with the results expressed as AUC;

[0085] (8) Based on the TCGA gastric cancer dataset, a nomogram model was constructed for BRCA1. By evaluating the clinical relevance of disease stage, age, gender, grade, and TMN stage, it was shown that there was a significant difference in age stage, and the calculated prognostic score showed a good prognosis.

[0086] (9) Based on the NHGRI-EBI Catalog of human genome-wide association studies (GWAS Catalog) and the IEU OpenGWAS database, the causal relationship between BRCA1, NADPH and gastric cancer was analyzed using the Mendelian randomization analysis method. The three were combined in pairs and it was found that BRCA1 mediates NADPH, destroys cellular oxidative stress, and disrupts cellular balance, thereby causing gastric cancer. Reverse Mendelian randomization analysis found that there was no causal relationship.

[0087] The present invention uses bioinformatics methods combined with Mendelian randomization methods to explore the mechanism of gastric cancer and realizes the mining of gastric cancer marker BRCA1. BRCA1 indirectly affects the occurrence of gastric cancer by mediating other proteins (NADPH), rather than directly intervening in the occurrence of gastric cancer through BRCA1. BRCA1 is a gastric cancer biomarker. BRCA1 uses NADPH oxidoreductase as an intermediary to destroy the redox homeostasis in the body, thereby causing the occurrence of gastric cancer.

[0088] Example 2

[0089] Based on the same principle as the method shown in Example 1 of the present invention, as shown in the attached Figure 2 As shown, an embodiment of the present invention further provides a gastric cancer biomarker mining system, including a data acquisition unit, a data preprocessing unit, a gene interaction network generation unit, a data difference analysis unit, a function enrichment analysis unit, a survival prognosis model gene analysis unit, a GSEA enrichment analysis unit, a progression-free survival analysis unit, a Nomogram model construction unit, a first causal relationship analysis unit, a second causal relationship analysis unit and a third causal relationship analysis unit;

[0090] A data acquisition unit is used to acquire gene expression data from the TCGA database, download clinical data of gastric cancer, and download data sets from the GEO database;

[0091] A data preprocessing unit, used to determine differentially expressed genes by setting differential expression thresholds, and obtain integrated gene expression data;

[0092] The gene interaction network generation unit is used to calculate the topological parameters of each node in the interaction network based on the integrated gene expression data by integrating the interaction network between genes, sort the topological parameters by size, screen them according to the set threshold, and determine the core gene target network;

[0093] Data difference analysis unit, used to perform core gene target network analysis based on the data set in the GEO database, analyze the differential expression of core genes in the intergroup relationship by identifying the differential expression of core genes in gastric cancer tissues and normal tissues, and evaluate the statistical difference by calculating the P value;

[0094] Functional enrichment analysis unit, used to perform overall GO and KEGG enrichment analysis on the core gene target network based on the statistical difference evaluation results, and to analyze the relationship between the core gene target network and individual core gene targets by evaluating the GO and KEGG enrichment of individual core genes;

[0095] The survival prognosis model gene analysis unit is used to compare the survival of core genes in normal tissues and gastric cancer tissues based on TCGA clinical survival data using the Kaplan-Meier method, and to determine the candidate key genes that affect the occurrence and survival of gastric cancer by combining statistical differences with functional enrichment analysis results, and to construct a survival prognosis model;

[0096] GSEA enrichment analysis unit, used to perform GSEA enrichment analysis on the survival prognosis model and calculate GSEA-GO and GSEA-KEGG results;

[0097] Progression-free survival analysis unit, used to perform progression-free survival analysis on the survival prognostic model based on TCGA clinical data, and to evaluate the specific parameters, sensitivity parameters, and specificity parameters of the survival prognostic model;

[0098] The Nomogram model building unit is used to build a Nomogram model for gastric cancer biomarkers based on the gastric cancer dataset of the TCGA database, evaluate the clinical relevance of the survival prognostic model in disease stage, age, gender, grade and TMN stage, and calculate the survival rate;

[0099] The first causal relationship analysis unit is used to calculate the causal relationship between gastric cancer biomarkers and gastric cancer based on the GWAS Catalog database and the IEU OpenGWAS database through the Mendelian randomization method, taking gastric cancer biomarkers as exposure factors and gastric cancer as outcome factors, and perform exchange verification;

[0100] The second causal relationship analysis unit was used to calculate the causal relationship between protease and gastric cancer by using the Mendelian randomization method, taking protease as the exposure factor and gastric cancer as the outcome factor, and to perform exchange verification;

[0101] The third causal relationship analysis unit is used to apply the Mendelian randomization method, taking the gastric cancer biomarker as the exposure factor and the protease as the outcome factor, calculate the causal relationship between the gastric cancer biomarker and the protease, and perform exchange verification.

[0102] Optionally, the integrated gene expression data are imported into the String database to calculate the mutual relationships, construct a PPI network to obtain the interaction network, and use the CytoHubba plug-in in the Cytoscape software to screen the core gene target network to obtain several core genes.

[0103] Optionally, during the functional enrichment analysis, the core gene target network is subjected to overall GO and KEGG enrichment analysis according to biological process, cellular component, molecular function and pathway.

[0104] Optionally, construct a survival prognosis model, construct a survival prognosis model, and calculate the risk score to determine the accuracy of the survival prognosis model, including: assuming that the risk score is risk score, i is the sample number, n is the total number of sample numbers, coefi is the score coefficient obtained by survival analysis, and id is the expression level of the gastric cancer biomarker gene, then:

[0105]

[0106] Optionally, five Mendelian algorithms, including MR Egger algorithm, weighted median algorithm, therandom-effects inverse variance-weighted algorithm, simple mode algorithm and weighted mode algorithm, are used to calculate the causal relationship between gastric cancer biomarkers and gastric cancer, the causal relationship between proteases and gastric cancer, and the causal relationship between gastric cancer biomarkers and proteases, respectively.

[0107] The above are only preferred embodiments of the present invention and are not intended to limit 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. A method for mining gastric cancer biomarkers, characterized in that: include: Data acquisition: Obtain gene expression data from the TCGA database, download gastric cancer clinical data, and download data sets from the GEO database; Data preprocessing: By setting the differential expression threshold, differentially expressed genes are determined to obtain integrated gene expression data; Gene interaction network generation: Based on the integrated gene expression data, by integrating the interaction network between genes, the topological parameters of each node in the interaction network are calculated, the topological parameters are sorted by size, and the core gene target network is determined based on the set threshold. Data difference analysis: Based on the data set in the GEO database, the core gene target network analysis was performed, and the differential expression of the core genes in the intergroup relationship was analyzed by identifying the differential expression of the core genes in gastric cancer tissues and normal tissues, and the statistical differences were evaluated by calculating the P value; Functional enrichment analysis: Based on the statistical difference evaluation results, the core gene target network was subjected to overall GO and KEGG enrichment analysis; Gene analysis of survival prognosis model: Based on TCGA clinical survival data, the Kaplan-Meier method was used to compare the survival of core genes in normal tissues and gastric cancer tissues. The statistical difference and functional enrichment analysis results were combined to determine the candidate key genes that affect the occurrence and survival of gastric cancer, and to construct a survival prognosis model. GSEA enrichment analysis: GSEA enrichment analysis was performed on the survival prognosis model, and GSEA-GO and GSEA-KEGG results were calculated; Progression-free survival analysis: Based on TCGA clinical data, the survival prognostic model was analyzed for progression-free survival, and the specific parameters, sensitivity parameters, and specificity parameters of the survival prognostic model were evaluated; Nomogram model construction: Based on the gastric cancer dataset of the TCGA database, a Nomogram model was constructed for gastric cancer biomarkers to evaluate the clinical relevance of the survival prognostic model in disease stage, age, gender, grade and TMN stage, and calculate the survival rate; Analysis of the causal relationship between gastric cancer biomarkers and gastric cancer: Based on the GWAS Catalog database and the IEUOpenGWAS database, the causal relationship between gastric cancer biomarkers and gastric cancer was calculated by Mendelian randomization method, with gastric cancer biomarkers as exposure factors and gastric cancer as outcome factors, and exchange verification was performed; Analysis of the causal relationship between protease and gastric cancer: Using the Mendelian randomization method, protease was used as the exposure factor and gastric cancer was used as the outcome factor to calculate the causal relationship between protease and gastric cancer, and exchange verification was performed; Analysis of the causal relationship between gastric cancer biomarkers and proteases: Using the Mendelian randomization method, BRCA1 was used as the exposure factor and proteases as the outcome factor to calculate the causal relationship between gastric cancer biomarkers and proteases, and exchange verification was performed.

2. A gastric cancer biomarker mining method according to claim 1, characterized in that: The integrated gene expression data were imported into the String database to calculate the relationships, construct the PPI network to obtain the interaction network, and use the CytoHubba plug-in in the Cytoscape software to screen the core gene target network to obtain several core genes.

3. A gastric cancer biomarker mining method according to claim 1, characterized in that: During the functional enrichment analysis, the core gene target network was subjected to overall GO and KEGG enrichment analysis according to biological process, cellular component, molecular function and pathway.

4. The method for mining gastric cancer biomarkers according to claim 1, characterized in that: Construct a survival prognosis model and calculate the risk score to determine the accuracy of the survival prognosis model, including: assuming that the risk score is risk score, i is the sample number, n is the total number of sample numbers, coefi is the score coefficient obtained by survival analysis, and id is the expression level of the gastric cancer biomarker gene, then:

5. The method for mining gastric cancer biomarkers according to claim 1, characterized in that: Five Mendelian algorithms, including MR Egger algorithm, weighted median algorithm, IVW algorithm, simple mode algorithm and weighted mode algorithm, were used to calculate the causal relationship between gastric cancer biomarkers and gastric cancer, the causal relationship between proteases and gastric cancer, and the causal relationship between gastric cancer biomarkers and proteases, respectively.

6. A gastric cancer biomarker mining method according to claim 1, characterized in that: Heterogeneity, sensitivity, and specificity were analyzed using the IVW algorithm and the MREgger algorithm.

7. A gastric cancer biomarker mining system, characterized in that: It includes data acquisition unit, data preprocessing unit, gene interaction network generation unit, data difference analysis unit, function enrichment analysis unit, survival prognosis model gene analysis unit, GSEA enrichment analysis unit, progression-free survival analysis unit, Nomogram model construction unit, first causal relationship analysis unit, second causal relationship analysis unit and third causal relationship analysis unit; A data acquisition unit is used to acquire gene expression data from the TCGA database, download clinical data of gastric cancer, and download data sets from the GEO database; A data preprocessing unit, used to determine differentially expressed genes by setting differential expression thresholds, and obtain integrated gene expression data; The gene interaction network generation unit is used to calculate the topological parameters of each node in the interaction network based on the integrated gene expression data by integrating the interaction network between genes, sort the topological parameters by size, screen them according to the set threshold, and determine the core gene target network; Data difference analysis unit, used to perform core gene target network analysis based on the data set in the GEO database, analyze the differential expression of core genes in the intergroup relationship by identifying the differential expression of core genes in gastric cancer tissues and normal tissues, and evaluate the statistical difference by calculating the P value; Functional enrichment analysis unit, used to perform overall GO and KEGG enrichment analysis on the core gene target network based on the statistical difference evaluation results, and to analyze the relationship between the core gene target network and individual core gene targets by evaluating the GO and KEGG enrichment of individual core genes; The survival prognosis model gene analysis unit is used to compare the survival of core genes in normal tissues and gastric cancer tissues based on TCGA clinical survival data using the Kaplan-Meier method, and to determine the candidate key genes that affect the occurrence and survival of gastric cancer by combining statistical differences with functional enrichment analysis results, and to construct a survival prognosis model; GSEA enrichment analysis unit, used to perform GSEA enrichment analysis on the survival prognosis model and calculate GSEA-GO and GSEA-KEGG results; Progression-free survival analysis unit, used to perform progression-free survival analysis on the survival prognostic model based on TCGA clinical data, and to evaluate the specific parameters, sensitivity parameters, and specificity parameters of the survival prognostic model; The Nomogram model building unit is used to build a Nomogram model for gastric cancer biomarkers based on the gastric cancer dataset of the TCGA database, evaluate the clinical relevance of the survival prognostic model in disease stage, age, gender, grade and TMN stage, and calculate the survival rate; The first causal relationship analysis unit is used to calculate the causal relationship between gastric cancer biomarkers and gastric cancer based on the GWAS Catalog database and the IEU OpenGWAS database through the Mendelian randomization method, taking gastric cancer biomarkers as exposure factors and gastric cancer as outcome factors, and perform exchange verification; The second causal relationship analysis unit was used to calculate the causal relationship between ADPH and gastric cancer by using the Mendelian randomization method, with protease as the exposure factor and gastric cancer as the outcome factor, and to perform exchange verification; The third causal relationship analysis unit is used to apply the Mendelian randomization method, taking the gastric cancer biomarker as the exposure factor and the protease as the outcome factor, calculate the causal relationship between the gastric cancer biomarker and the protease, and perform exchange verification.

Images