Systematic prediction and layered screening method of renal cell carcinoma biomarkers

By constructing a renal cell carcinoma prognosis model based on race, pathological staging and lymph node metastasis characteristics, differentially expressed mRNA and miRNA were screened out, and a risk scoring system was established, which solved the problem of lack of precise biomarkers in the existing technology, and achieved the goal of High reliability prognosis prediction and diagnosis in patients with renal cell carcinoma.

CN119020488AInactive Publication Date: 2024-11-26GUANGDONG MEDICAL UNIV
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410882139.7
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-03
Publication Date
2024-11-26
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

The prior art lacks precise biomarkers and stratified screening methods for different human races and tumor stages, making it difficult to effectively predict and diagnose early lesions of renal cell carcinoma.

Method used

The prognosis model of renal cell carcinoma based on the characteristics of race, pathological staging and lymph node metastasis was constructed. Through multi-factor Cox regression analysis and Lasso analysis, differentially expressed mRNA and miRNA were screened out, and mRNA, miRNA or combined mRNA+miRNA prognosis model was established. Risk scores and screen biomarkers in stratified.

Benefits of technology

It has achieved high-reliability prognosis prediction for patients with renal cell carcinoma in different human races and tumor stages, provided accurate diagnostic and targeted treatment references, and improved the early identification and prognosis evaluation capabilities of patients with renal cell carcinoma.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119020488A_ABST
    Figure CN119020488A_ABST
Patent Text Reader

Abstract

The invention belongs to the technical field of medical information, and particularly relates to a systematic prediction and layered screening method of renal cell carcinoma biomarkers. The invention provides 35 renal cell carcinoma biomarkers which comprise 17 renal cell carcinoma prognosis genes, a layered screening method of the biomarkers is to perform screening and identification by constructing prognosis models, and three prognosis models, namely mRNA prognosis models, mRNA prognosis models, mRNA prognosis models, mRNA prognosis models, mRNA prognosis models, mRNA prognosis models, mRNA prognosis models, mRNA prognosis models, mRNA prognosis models, mRNA prognosis models, mRNA prognosis models, mRNA prognosis models, mRNA prognosis models and mRNA prognosis models each type of the mRNA prognosis model and the mRNA + miRNA combined prognosis model is divided into a first-level RCC group, a second-level white human, a Stage I-II stage, a Stage III-IV stage and a lymphatic non-metastasis group, a third-level white human Stage I-II stage, a white human Stage III-IV stage and a white human lymphatic non-metastasis group, and the total number of the three prognosis models is 24. The reliability of each group of models on prognosis prediction of RCC patients is high (the AUC values in three years and five years are basically 0.7 or above). In different prognosis models, 45 kinds of mRNA biomarkers and 33 kinds of miRNA biomarkers are identified in total.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the field of medical information technology, and specifically relates to a method for systematically predicting and stratifying biomarkers of renal cell carcinoma. Background Art

[0002] Renal cell carcinoma (RCC), also known as renal adenocarcinoma, is one of the top ten cancers that affect human survival, accounting for 90% of renal malignant tumors and one of the most deadly malignant tumors of the urinary system. Existing studies have found that different races and tumor stages have different effects on the occurrence and development of RCC. The incidence of renal cancer in different races is also different. The incidence of renal cell carcinoma in blacks is much higher than that in whites. Tumor pathological stage is also a major factor affecting the survival of renal cancer patients. In addition, lymph node metastasis can impair the prognosis of renal cancer. Therefore, it is crucial to study the effects of various clinical variables, including race, pathological stage, and lymph node metastasis, on the gene expression of RCC patients. This correlation is extremely meaningful for the early diagnosis and comprehensive evaluation of people at high risk of renal cancer.

[0003] Messenger RNA (mRNA) plays an important role in the occurrence and development of cancer. Molecular pathways in cells can change the expression of mRNA, thereby affecting the translation of mRNA into protein. The dysregulation of microRNA (miRNA) expression is also related to various cancers. By affecting the signal transduction pathways in cellular processes, it binds to molecules such as mRNA and regulates its physiological effects. Changes in the expression of miRNA and mRNA in cells play a key role in the invasion process of tumors, so these RNAs can be regarded as biomarkers for diagnosis and prognosis.

[0004] At present, the specific mechanism of cancer development is still unclear, and considering the changes in race, tumor stage and metastasis status, the existing technology lacks accurate biomarkers for renal cell carcinoma under various pathological conditions and methods for stratified screening of renal cell carcinoma markers. Summary of the invention

[0005] The object of the present invention is to provide a systematic renal cell carcinoma biomarker and a stratified screening method thereof.

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

[0007] A renal cell carcinoma biomarker, wherein the marker is one or more of the following:

[0008]

[0009]

[0010] Among them, Δ represents the up-regulation of DEGs-prognostic or DEMs-prognostic predicted in this study, Represents the down-regulation of DEGs-prognostic or DEMs-prognostic predicted in this study.

[0011] The gene IDs in the table are the codes in the NCBI database and are uniquely identified.

[0012] The Accession in the table is queried through the mirBase database and refers to the mirRNA precursor number, which is the only definite code.

[0013] The present invention includes prognostic genes for renal cell carcinoma, which are one or more of 17 prognostic genes, including 14 upregulated prognostic genes (CCDC194, BORCS8-MEF2B, CST2, CDKN2C, ZIC5, PCDHGC5, ERICH6B, ZACN, MXRA8, RPP21, COL6A2, DEPP1, C1orf162, PLEKHG4) and 3 downregulated prognostic genes (TMEM200C, FOXE3, IL13RA2).

[0014] A method for screening renal cell carcinoma biomarkers, wherein the screening method is to screen through a renal cell carcinoma prognostic model, wherein the screening specifically constructs a stratified prognostic model based on a characteristic system of race, pathological stage, and whether lymph node metastasis occurs, and after evaluating the prognostic model, obtains the renal cell carcinoma biomarkers included in the model;

[0015] The method for constructing the renal cell carcinoma prognosis model comprises the following steps:

[0016] (1) Obtain renal cell carcinoma-related mRNA and miRNA, group them, analyze them, and optimize them;

[0017] (2) Establish a prognostic model by multivariate Cox regression analysis;

[0018] (3) Evaluate the prognostic model.

[0019] Preferably, the grouping is to divide the samples into different groups according to the characteristics of race, pathological stage and whether lymph nodes are metastatic. These groups include total RCC samples, white samples, black samples, Stage I-II samples, Stage III-IV samples, white stage I-II samples, white stage III-IV samples, black stage I-II samples, black stage III-IV samples, no lymph node metastasis samples, lymph node metastasis samples, white no lymph node metastasis samples, white lymph node metastasis samples and black lymph node metastasis samples.

[0020] Preferably, the analysis is: using the R package "edgeR" to perform differential expression analysis, screening out differentially expressed mRNAs and miRNAs, and naming them DEGs and DEMs, respectively.

[0021] Preferably, the analysis further comprises: performing univariate Cox analysis and KM analysis on DEGs and DEMs, and taking the intersection of the two sets of analysis results to obtain DEGs-os and DEMs-os, respectively.

[0022] Preferably, the optimization is: optimizing DEGs-os and DEMs-os respectively through Lasso analysis, and merging the optimized DEGs-os and DEMs-os matrices to obtain DEGs-os+DEMs-os.

[0023] Preferably, the prognostic model comprises mRNA, miRNA or mRNA+miRNA prognostic model.

[0024] Preferably, the method for establishing a prognostic model is: using the "survminer" package to perform multivariate Cox proportional hazard regression analysis on DEGs-os, DEMs-os and DEGs-os+DEMs-os markers and construct a prognostic model.

[0025] Preferably, evaluating the prognostic model includes: calculating the risk score of renal cell carcinoma patients according to the expression of prognostic DEGs-os, prognostic DEMs-os and prognostic DEGs-os+DEMs-os, and the calculation formula is as follows:

[0026] Risk Score=Exp(DEG1 / DEM1)×β1+Exp(DEG2 / DEM2)×β2+...+Exp(DEGn / DEMn)×βn,

[0027] Where: Exp(DEG / DEM) represents the expression of prognostic DEG-os / DEM-os, and β represents the coefficient of the prognostic DEG-os / DEM-os. The β coefficient is determined by the specific marker and is analyzed by the survival and survminer R packages of R software.

[0028] Preferably, evaluating the prognostic model includes: dividing RCC patients into high-risk group and low-risk group by median risk score, drawing risk score curve and expression heat map, and drawing survival curves of patients in high-risk and low-risk groups by KM analysis.

[0029] Preferably, evaluating the prognostic model includes: drawing ROC curves of each of the mRNA, miRNA and mRNA+miRNA prognostic models to evaluate their predictive abilities.

[0030] Preferably, evaluating the prognostic model includes: using Spearman correlation analysis to evaluate the relationship between the stratification depth of the RCC prognostic model and the three-year and five-year survival AUC values.

[0031] Preferably, the method for constructing a prognostic model further comprises: using literature retrieval to check the prognostic mRNA and miRNA obtained, and comparing their expression with the corresponding prognostic RNA in existing RCC-related literature; to verify the credibility of the prognostic model of this study and identify new molecular biomarkers.

[0032] Beneficial Effects

[0033] The present invention constructs three prognostic models for patients with renal adenocarcinoma of different races and stages, namely, mRNA prognostic model, miRNA prognostic model and mRNA+miRNA combined prognostic model, each type is divided into the first level RCC group, the second level white, Stage I-II, Stage III-IV and lymph node non-metastasis group, the third level white Stage I-II, white Stage III-IV and white lymph node non-metastasis group, a total of 8 groups, three prognostic models a total of 24 groups. The reliability of each group of models for the prognosis prediction of RCC patients is high (three-year and five-year AUC values ​​are basically above 0.7).

[0034] The present invention predicts mRNAs and miRNAs with prognostic significance for RCC patients under corresponding stratification, providing valuable reference for accurate diagnosis and targeted treatment of RCC patients. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0036] Figure 1 The flowchart of the study.

[0037] Figure 2 Figure 2 shows the risk heat map and risk curve of the mRNA prognostic model.

[0038] Figure 3 Figure 2 shows the risk heat map and risk curve of the miRNA prognostic model.

[0039] Figure 4 This is the risk heat map and risk curve of the mRNA+miRNA combined prognostic model.

[0040] Note: Figures 2 to 4 In the figure, A is the RCC group, B1 is the white group, C1 is the Stage I-II group, C2 is the Stage III-IV group, D1 is the white Stage I-II group, D2 is the white Stage III-IV group, E1 is the lymph node non-metastasis group, and F1 is the white lymph node non-metastasis group. The heat map of the prognostic gene expression profiles of the corresponding groups and the distribution of the risk scores of RCC patients are presented from top to bottom.

[0041] Figure 5 Figure 2 shows the Kaplan-Meier curves of the low-risk and high-risk groups of patients in the mRNA prognostic model.

[0042] Figure 6 Figure 2 shows the Kaplan-Meier curves of the low-risk and high-risk groups of patients in the miRNA prognostic model.

[0043] Figure 7 Kaplan-Meier curves of low-risk and high-risk groups of patients in the mRNA+miRNA combined prognostic model.

[0044] Note: Figures 5 to 7 In the figure, A is the RCC group, B1 is the white group, C1 is the Stage I-II group, C2 is the Stage III-IV group, D1 is the white Stage I-II group, D2 is the white Stage III-IV group, E1 is the lymph node non-metastasis group, and F1 is the white lymph node non-metastasis group. The horizontal axis represents the survival time of RCC patients in the corresponding group, and the vertical axis represents the 5-year survival rate. The blue curve represents the 5-year survival rate of RCC patients in the low-risk group, and the red curve represents the 5-year survival rate of RCC patients in the high-risk group.

[0045] Figure 8 ROC curve of the mRNA prognostic model based on risk score for predicting the overall survival of RCC patients.

[0046] Fig. 9 ROC curve of the miRNA prognostic model based on risk score for predicting the overall survival of RCC patients.

[0047] Fig.10 ROC curve of the mRNA+miRNA combined prognostic model based on risk score for predicting the overall survival of RCC patients.

[0048] Note: Figures 8 to 10 In the figure, A is the RCC group, B1 is the white group, C1 is the Stage I-II group, C2 is the Stage III-IV group, D1 is the white Stage I-II group, D2 is the white Stage III-IV group, E1 is the lymph node non-metastasis group, and F1 is the white lymph node non-metastasis group. The horizontal axis represents the false positive rate (1-specificity) of the prognostic model predicting the overall survival of each group of RCC patients. The closer it is to zero, the higher the accuracy; the vertical axis represents the true positive rate (sensitivity) of the prognostic model predicting the overall survival of each group of RCC patients. The larger the sensitivity, the higher the accuracy. The blue line is five-year survival, and the red line is three-year survival. The larger the area under the curve, that is, the higher the AUC value, the higher the accuracy of the prognostic model in predicting the overall survival of each group of RCC patients.

[0049] Fig.11 A is the AUC flow chart. AC is the three-year and five-year survival AUC value flow chart of the ROC curve of the mRNA prognostic model, miRNA prognostic model, and mRNA+miRNA prognostic model.

[0050] Fig.12 is the Spearman coefficient correlation analysis. (A)(C)(E) is the Spearman coefficient correlation analysis used to determine the correlation between the stratification depth and the three-year survival AUC value. (B)(D)(F) is the Spearman coefficient correlation analysis used to determine the correlation between the stratification depth and the five-year survival AUC value. Among them, (A) and (B) represent the mRNA prognostic model, (C) and (D) represent the miRNA prognostic model, and (E) and (F) represent the mRNA+miRNA combined prognostic model.

[0051] Fig.13 Flowchart of literature search for eligible studies. DETAILED DESCRIPTION

[0052] The present invention provides a method for constructing a prognostic model for renal cell carcinoma. The construction process is as follows: Figure 1 shown.

[0053] Figure 1 The present invention shows the process of constructing a prognostic model for RCC by stratification, exploring the correlation between model stratification and accurate prognosis, and mining new biomarkers. Specifically, the present invention is divided into the following six steps:

[0054] (1) Download RCC-related mRNA and miRNA expression profiles from the TCGA database. Then, the samples were divided into different groups according to characteristics such as race, pathological stage, and lymph node metastasis. Therefore, we selected a total of 8 groups that met the requirements, namely, RCC group (A), white group (B1), Stage I-II group (C1), Stage III-IV group (C2), white stage I-II group (D1), white stage III-IV group (D2), no lymph node metastasis group (E1), and white no lymph node metastasis group (F1);

[0055] (2) Differential expression analysis was performed using the R package “edgeR”, and eight groups of DEGs and DEMs were obtained;

[0056] (3) Univariate Cox analysis and KM analysis were performed on the eight groups of DEGs and DEMs. The two analyses used P lower than 0.01 and 0.05 as screening conditions for DEGs and DEMs, respectively. The results of the two analyses were intersected to obtain DEGs-os and DEMs-os, respectively;

[0057] (4) Eight groups of DEGs-os and DEMs-os were optimized by Lasso analysis, and the eight optimized DEGs-os and DEMs-os matrices were merged to obtain eight groups of DEGs-os + DEMs-os;

[0058] (5) Eight groups of mRNA, miRNA, and mRNA+miRNA prognostic models were established by multivariate Cox regression analysis, and the risk heat map, risk curve, survival curve, and ROC curve of each group were drawn. Spearman correlation analysis was then used to evaluate the relationship between the stratification depth of the RCC prognostic model and the three- and five-year survival AUC values;

[0059] (6) Use literature search to examine the prognostic mRNAs (DEGs-prognostic) and miRNAs (DEMs-prognostic) obtained in this study, and compare their expression with the corresponding prognostic RNAs in existing RCC-related literature to verify the credibility of the prognostic model of this study and identify new molecular biomarkers;

[0060] All statistical analyses in the examples of the present invention were based on R 4.1.2. If not specified, a P value lower than 0.05 was considered statistically significant, and all P values ​​were two-sided.

[0061] In order to further illustrate the present invention, the technical solution provided by the present invention is described in detail below in conjunction with the accompanying drawings and embodiments, but they should not be construed as limiting the protection scope of the present invention.

[0062] Example 1

[0063] Acquisition, grouping, prediction and selection of RCC mRNA and miRNA

[0064] 1. Acquisition of RCC mRNA and miRNA datasets and clinical data

[0065] A total of 991 RCC samples, including 864 tumor samples and 127 normal samples, were downloaded from The Cancer Genome Atlas (TCGA) database (https: / / portal.gdc.cancer.gov / ). Subsequently, the relevant clinical data of RCC in different ethnic groups and stages were downloaded.

[0066] 2. Sample grouping of RCC mRNA and miRNA datasets

[0067] Multiple groups of mRNA and miRNA data were retrieved from the TCGA database. Specifically, these groups included total RCC samples, white samples, black samples, stage I-II samples, stage III-IV samples, white stage I-II samples, white stage III-IV samples, black stage I-II samples, black stage III-IV samples, no lymph node metastasis samples, lymph node metastasis samples, white no lymph node metastasis samples, white lymph node metastasis samples, and black lymph node metastasis samples (see Table 1). To ensure the accuracy of the results, 7 groups with less than 20 tumor samples or normal samples were excluded, including the black group (B2), black stage I-II group (D3), black stage III-IV group (D4), lymph node metastasis group (E2), white lymph node metastasis group (F2), black no lymph node metastasis group (G1), and black lymph node metastasis group (G2). Finally, the 8 groups that met the requirements in the above table were included in the subsequent analysis.

[0068] Table 1 Number of RCC samples

[0069]

[0070]

[0071] Note: The bold groups are those included in the analysis.

[0072] 3. Evaluation of differential expression of mRNA and miRNA

[0073] The “edgeR” package of R software was used to analyze the expression matrix of eight groups of mRNAs and miRNAs from the TCGA database, and then the default Benjamini-Hochberg false discovery rate (FDR) method was used to screen out differentially expressed mRNAs (DEGs) and miRNAs (DEMs) with FDR < 0.05 and |log2FC| > 1 as the screening conditions.

[0074] Prediction of RCC DEGs and DEMs

[0075] Eight groups of mRNAs and miRNAs with statistically significant differential expression were obtained from the TCGA database, namely, DEGs and DEMs contained 5078, 5112, 5072, 5334, 5056, 5271, 5244, 5307 mRNAs and 221, 228, 223, 241, 237, 244, 211, 196 miRNAs, respectively, which were identified between tumor and normal tissues.

[0076] 4. Univariate analysis and survival analysis based on DEGs and DEMs

[0077] The 8 groups of DEGs and DEMs were subjected to univariate Cox proportional hazard regression analysis and Kaplan-Meier (KM) survival analysis using the "survival" package of R language, thus obtaining 8 groups of DEGs-uc / DEMs-uc (DEGs / DEMs-Univariate Cox) and DEGs-km / DEMs-km (DEGs / DEMs-KM), respectively. Both analysis methods used P<0.01 as the screening condition for mRNAs and P<0.05 as the screening condition for miRNAs, and the intersection of the univariate Cox regression analysis results and the KM analysis results was taken to obtain the DEGs (DEGs-os) and DEMs (DEMs-os) significantly associated with the overall survival of RCC patients.

[0078] Selection of DEGs-os and DEMs-os

[0079] Table 2 shows the number of DEGs-uc and DEMs-uc obtained by univariate Cox analysis, and the number of DEGs-km and DEMs-km obtained by KM analysis. Combining the results of the two groups of analyses, the number of DEGs-os was 2180, 2114, 672, 962, 531, 981, 1890, and 1908, and the number of DEMs-os was 100, 92, 31, 39, 30, 46, 69, and 62, respectively, in the RCC group, white group, stage I-II group, stage III-IV group, white stage I-II, white stage III-IV, no lymph node metastasis group, and white no lymph node metastasis group. In addition, these DEGs-os and DEMs-os were used to construct the subsequent prognostic model.

[0080] Table 2 Number of DEGs-uc, DEMs-uc, DEGs-km, DEMs-km, DEGs-os, and DEMs-os in each group

[0081]

[0082] Example 2

[0083] Construction of a prognostic model for renal cell carcinoma

[0084] Three prognostic models were established for 8 groups of DEGs-os, 8 groups of DEMs-os and 8 groups of DEMs-os+DEGs-os, namely mRNA, miRNA and mRNA+miRNA prognostic models.

[0085] To reduce the risk of overfitting, the lasso Cox regression analysis method was applied, i.e., variable selection and shrinkage were performed through the R language "glmnet" package. The optimal value of the penalty parameter lambda was determined through 300 cross-validations based on the minimum criterion, and a series of markers with corresponding coefficients were obtained.

[0086] The matrices of the eight groups of DEGs-os and DEMs-os after Lasso analysis were merged to obtain eight groups of DEGs-os + DEMs-os.

[0087] Then, the “survminer” package was used to perform multivariate Cox proportional hazard regression analysis on DEGs-os, DEMs-os, and DEGs-os+DEMs-os markers and construct a prognostic model.

[0088] The risk score of each patient was calculated based on the expression of prognostic DEGs-os (DEGs-prognostic), prognostic DEMs-os (DEMs-prognostic), and prognostic DEGs-os + DEMs-os (DEGs-prognostic + DEMs-prognostic), and the calculation formula was as follows:

[0089] Risk Score=Exp(DEG1 / DEM1)×β1+Exp(DEG2 / DEM2)×β2+...+Exp(DEGn / DEMn)×βn,

[0090] Where: Exp(DEG / DEM) represents the expression level of prognostic DEG-os / DEM-os, and β represents the coefficient of the prognostic DEG-os / DEM-os.

[0091] RCC patients were divided into high-risk and low-risk groups by median risk score, and risk score curves and expression heat maps were drawn. Survival curves of patients in high-risk and low-risk groups were drawn by KM analysis. ROC curves of eight groups of prognostic models, including mRNA, miRNA, and mRNA+miRNA, were then drawn to evaluate their predictive ability, and the spearman correlation coefficient was used to evaluate the correlation between RCC stratification depth and three-year and five-year survival AUC values.

[0092] In this embodiment, mRNA and miRNA prognostic models were constructed by LASSO regression analysis and multifactor Cox regression analysis, respectively. Table 3 shows the markers of 8 groups of mRNA prognostic models, including 38 DEGs-prognostics, and Table 4 shows the markers of 8 groups of miRNA prognostic models, including 31 DEMs-prognostics. DEGs-os and DEMs-os after LASSO analysis were combined for multifactor Cox regression analysis, and 8 groups of mRNA+miRNA joint prognostic models were established. Table 5 shows the markers of each group of mRNA+miRNA prognostic models, and 59 DEGs-prognostic+DEMs-prognostics were determined, including 34 DEGs-prognostics and 25 DEMs-prognostics. Among all the above-mentioned predicted biomarkers, there are 78 biomarkers, including 45 mRNAs and 33 miRNAs.

[0093] Table 3 Prognostic markers of eight groups of mRNA models

[0094]

[0095]

[0096] Note: The * indicates the same mRNA in each group.

[0097] Table 4 Prognostic markers of eight groups of miRNA models.

[0098]

[0099]

[0100] Note: The * indicates the same miRNA in each group.

[0101] Table 5 Prognostic markers of eight groups of mRNA+miRNA models

[0102]

[0103]

[0104]

[0105] Note: The * indicates the same mRNA and miRNA in each group.

[0106] Evaluation of the predictive ability of prognostic models

[0107] The risk heat maps and risk curves of the mRNA, miRNA, and mRNA+miRNA prognostic models are shown in Figure 2 . Figure 2-4 As shown. The risk heat map showed that in the mRNA, miRNA and mRNA+miRNA prognostic models, there were significant differences in the expression of DEGs-prognostic, DEMs-prognostic and DEGs-prognostic+DEMs-prognostic between the high-risk group and the low-risk group. In the 24 risk curves drawn by the above three prognostic models, the risk scores of patients in the low-risk group were small and consistent, while the risk scores of patients in the high-risk group were large and significantly increased.

[0108] The Kaplan-Meier survival curve showed that the survival rate of patients in the low-risk group was significantly higher than that in the high-risk group in all 24 prognostic models, and the difference was statistically significant (P<0.05). Figure 5-7 shown.

[0109] The three-year survival AUC values ​​of the eight mRNA prognostic models were basically above 0.74. Except for the stage I-II group (three-year AUC = 0.71), the five-year survival AUC values ​​were all greater than 0.75 (see Figure 8However, the three-year survival AUC values ​​of the eight miRNA prognostic models were basically above 0.70. Except for the Stage III-IV group (three-year AUC = 0.68), the five-year AUC values ​​were all above 0.72 (see Fig. 9 ). Similarly, the three-year AUC values ​​of the eight groups of mRNA+miRNA prognostic models were basically above 0.75, except for the Stage I-II group (three-year AUC = 0.74), and the five-year AUC values ​​were all above 0.80 (see Fig.10 In addition, except for the Stage III-IV group of the miRNA prognostic model, the AUC values ​​of the three-year survival and five-year survival of the other 23 prognostic models were all greater than 0.7, indicating that the models have the potential to reasonably predict the prognosis of patients. The five-year survival AUC of the 24 prognostic models is higher than the three-year survival AUC, suggesting that the model is better at predicting the five-year survival of RCC patients than the three-year survival.

[0110] The AUC flow chart of the prediction model of the present invention shows that as the model layer depth increases, the AUC value gradually increases, indicating that its prediction ability gradually increases (see Fig.11 ). This trend was more obvious in the stratified comparison of the five-year AUC values ​​of the mRNA+miRNA combined prognostic model. The AUC value of the first layer (RCC group) was 0.81, the AUC value of the second layer (white, Stage I-II, Stage III-IV, lymph node non-metastasis group) was greater than 0.80, and the AUC value of the third layer (white Stage I-II, white Stage III-IV, white lymph node non-metastasis group) was greater than 0.84. It is worth noting that compared with the corresponding groups of the single mRNA or miRNA prognostic model, the AUC value of each group of the mRNA+miRNA combined prognostic model increased, suggesting that the mRNA+miRNA prognostic model is more accurate and effective than the mRNA and miRNA prognostic models in predicting the prognosis of RCC.

[0111] The results of Spearman coefficient correlation analysis showed that the stratification levels of the three prognostic models established in this study were positively correlated with the three-year and five-year survival AUC values ​​(rs>0) (see Fig.12 Among them, the linear positive correlation between the number of stratification and the five-year survival AUC value of the mRNA+miRNA prognostic model was statistically significant (P<0.05) (see Fig.12 F), that is, there is a gradually increasing linear trend between the stratification depth and the AUC value of the prognostic model.

[0112] Example 3

[0113] Literature verification

[0114] According to the obtained prognostic DEGs-os and DEMs-os, the advanced search function of databases such as PubMed (https: / / pubmed.ncbi.nlm.nih.gov / ), Embase (https: / / www.embase.com / ), and CNKI (China National Knowledge Infrastructure, https: / / www.cnki.net / ) was used to enter keywords such as "prognostic mRNAs or miRNAs" and "renal cell carcinoma" to retrieve relevant literature studies published on or before February 28, 2024. In addition, we also compared the abnormal changes in the expression of prognostic markers of DEGs and DEMs with those reported in previous RCC-related literature.

[0115] The present invention systematically conducted a literature search to explain the relationship between predicted prognostic markers and RCC (see Fig.13 ). As shown in Tables 6-7, the abnormal expression of DEGs-prognostic, DEMs-prognostic and DEGs-prognostic+DEMs-prognostic in 8 groups of mRNA, miRNA and mRNA+miRNA prognostic models was compared with the abnormal expression of corresponding RNAs in previous RCC-related literature. In short, 19 of the 45 abnormally expressed DEGs-prognostic obtained in each group of RCC prognostic models were confirmed in early experiments, while HJURP showed the opposite abnormal expression pattern in the wet experiment. In addition, 9 DEGs-prognostic have been verified by previous bioinformatics studies, confirming the robustness of the results of the present invention. Seventeen novel prognostic genes were discovered through analysis, including 14 upregulated prognostic genes (CCDC194, BORCS8-MEF2B, CST2, CDKN2C, ZIC5, PCDHGC5, ERICH6B, ZACN, MXRA8, RPP21, COL6A2, DEPP1, C1orf162, PLEKHG4) and 3 downregulated prognostic genes (TMEM200C, FOXE3, IL13RA2).

[0116] Among the 33 predicted DEMs-prognostic molecular biomarkers, the abnormal expression levels of 9 in RCC were consistent with the retrieved experimental research results. The abnormal expression of 6 DEMs-prognostics in RCC was consistent with the bioinformatics analysis results of others. The remaining 18 new DEMs-prognostic molecular biomarkers for RCC, including 13 upregulated DEMs-prognostics (hsa-miR-3922-3p, hsa-miR-7702, hsa-miR-301b-3p, hsa-miR-452-3p, hsa-miR-3677-5p, hsa-miR-34a-3p, hsa-miR-6718-5p, hsa-miR-576-5p, hsa-miR-590-3p, hsa-miR-4732-3p, hsa-miR-5586-5p, hsa-miR-1307-5p, hsa-miR-1228-3p) and 5 down-regulated DEMs-prognostic (hsa-miR-3065-5p, hsa-miR-376b-3p, hsa-miR-369-5p, hsa-miR-10b-3p, hsa-miR-376b-5p).

[0117] Table 6 Abnormal expression of DEGs-prognostic in this study compared with previous RCC literature.

[0118]

[0119]

[0120]

[0121] Note: Representative experiments reported that DEGs-prognostic were upregulated in RCC, which was consistent with our findings. Representative experiments reported that DEGs-prognostic were down-regulated in RCC, which was consistent with our findings.

[0122] Representative experiments reported that DEGs-prognostic were down-regulated in RCC, which was inconsistent with our results.

[0123] Representative computationally reported DEGs-prognostic upregulation in RCC, which was consistent with our results.

[0124] Table 7 Abnormal expression of DEMs-prognostic in this study compared with previous RCC literature.

[0125]

[0126]

[0127] Note: Representative experiments reported that DEMs-prognostic upregulation in RCC was consistent with our findings.

[0128] Representative experiments reported that DEMs-prognostic were down-regulated in RCC, which was consistent with our findings.

[0129] Representative computationally reported DEMs-prognostic upregulation in RCC, which is consistent with our results.

[0130] Representative computationally reported DEMs-prognostic downregulation in RCC, which is consistent with our results.

[0131] It can be seen from the above embodiments that the present invention constructs three prognostic models for patients with renal adenocarcinoma of different races and stages in a hierarchical manner, namely, an mRNA prognostic model, a miRNA prognostic model, and an mRNA+miRNA combined prognostic model. Each type is divided into a first-level RCC group, a second-level white patient, Stage I-II, Stage III-IV, and a lymph node non-metastasis group, a third-level white patient Stage I-II, white patient Stage III-IV, and white patient lymph node non-metastasis group, a total of 8 groups, and a total of 24 groups of three prognostic models.

[0132] The present invention obtained 78 prognostic molecular biomarkers from 24 groups of prognostic models, including 45 DEGs-prognostics and 33 DEMs-prognostics, of which 28 DEGs-prognostics (Table 6) and 15 DEMs-prognostics (Table 7) were reported in the relevant literature of RCC. In addition, RCC-related experimental studies reported 21 upregulated prognostic molecular markers (15 DEGs-prognostics and 6 DEMs-prognostics) and 8 downregulated prognostic molecular markers (5 DEGs-prognostics and 3 DEMs-prognostics), of which 28 DEGs / DEMs-prognostics were consistent with the retrieved experimental results, accounting for 35.90% of all predictive biomarkers. However, the prediction of abnormal expression of a DEGs-prognostic (HJURP) in the present invention was inconsistent with the retrieved experimental results. According to our prediction, HJURP is upregulated in RCC, while previous experiments showed that its expression is downregulated, which may be caused by racial differences. Because our samples are white RCC in the TCGA database, Yuan et al.

[20] used East Asian RCC samples obtained from the Affiliated Hospital of Qingdao University. In addition, studies using the TCGA database for bioinformatics analysis of RCC have obtained consistent results with ours [21, 22], so it can ultimately be attributed to consistency. Similarly, in this study, the abnormal expression of 8 DEGs-prognostics (8 all upregulated) and 6 DEMs-prognostics (4 upregulated and 2 downregulated) was consistent with other bioinformatics analysis results. Therefore, the expression of 55.13% (43 out of 78) of the biomarkers predicted in this study is consistent with the results reported in the literature, and the remaining 44.87% (35 out of 78) are new biomarkers (Table 8). Therefore, the prognostic markers predicted in the present invention are significantly consistent with the existing findings in the literature. This consistency emphasizes the robustness and reliability of the prognostic models established for each specific group in our analysis. The comprehensive approach of the present invention helps to improve the overall confidence in the predictive power of these prognostic models, taking into account various factors and stages. The consistency between our predictions and existing knowledge further strengthens the validity and utility of our findings in mining prognostic biomarkers for renal cell carcinoma.

[0133] Table 8 Novel biomarkers discovered by the present invention

[0134]

[0135]

[0136] Note: Δ represents the up-regulation of DEGs-prognostic or DEMs-prognostic predicted by this study and not previously reported in RCC.

[0137] The downregulation of DEGs-prognostic or DEMs-prognostic predicted in this study has not been previously reported in RCC.

[0138] The present invention systematically constructs a stratified prognostic model for RCC based on characteristics such as race, stage, and lymph node metastasis status, establishes 24 groups of new prognostic models of different categories, and improves the prediction accuracy of the model. At the same time, the present invention also obtains 35 new molecular biomarkers for RCC prognosis. In addition, with the increase in stratification depth, the prognostic model has improved in predicting the prognosis of RCC, providing valuable insights for strengthening the identification of different RCC high-risk populations, clinical decision-making, and personalized precision medicine, which helps to improve the prognosis of patients.

[0139] Although the above embodiment describes the present invention in detail, it is only a part of the embodiments of the present invention, not all of the embodiments. People can also obtain other embodiments based on this embodiment without creativity, and these embodiments all fall within the protection scope of the present invention.

[0140] Appendix: The references cited in the specification of the present invention and Tables 6-7 are:

[0141] [1]S.Zhang,K.Xia,Y.Chang,Y.Wei,Y.Xiong,F.Tang,and J.Peng,LRP2 andDOCK8 Are Potential Antigens for mRNAVaccine Development in Immunologically'Cold'KIRC Tumours,Vaccines(Basel),vol.11,no.2,p.396,2023.

[0142] [2] Y. Wang, Y. Chen, B. Zhu, L. Ma, and Q.

[0143] [3]S.Liu,Y.Wang,C.Miao,Q.Xing,andZ.Wang,High expression ofCDCA7predicts poorprognosis for clear cell renal cell carcinoma andexplores itsassociations with immunity,Cancer Cell Int.,vol.21,no.1,p.140,2021.

[0144] [4]H.Ito,O.Ichiyanagi,S.Naito,V.N.Bilim,Y.Tomita,T.Kato,A.Nagaoka,andN.Tsuchiya,GSK-3directly regulates phospho-4EBP1 in renal cell carcinomacell-line:an intrinsic subcellular mechanism for resistance to mTORC1inhibition,BMC Cancer,vol.16,p.393,2016.

[0145] [5]W.Han,B.Fan,Y.Huang,X.Wang,Z.Zhang,G.Gu,and Z.Liu,Construction andvalidation of a prognostic model ofRNA binding proteins in clear cell renalcarcinoma,BMC Nephrol.,vol.23,no.1,p.172,2022.

[0146] [6]Y.Qian,Y.Li,K.Chen,N.Liu,X.Hong,D.Wu,Z.Xu,L.Zhou,L.Xu,and R.Jia,etal.,Pan-Cancer TranscriptomicAnalysis Identifies PLK1 Crucial for theTumorigenesis ofClear Cell Renal Cell Carcinoma,J.Inflamm.Res.,vol.15,pp.1099-1116,2022.

[0147] [7]W.Li,X.Meng,H.Yuan,W.Xiao,and X.Zhang,M2-polarization-relatedCNTNAP1 gene might be a novel immunotherapeutic target andbiomarker for clearcell renal cell carcinoma,IUBMB Life,vol.74,no.5,pp.391-407,2022.

[0148] [8]S.Liu,Y.Yu,Y.Wang,B.Zhu,and B.Han,COLGALT1 is a potentialbiomarker for predicting prognosis and immune responses for kidney renalclear cell carcinoma and its mechanisms of ceRNA networks,Eur.J.Med.Res.,vol.27,no.1,p.122,2022.

[0149] [9]G.Sun,Y.Ge,Y.Zhang,L.Yan,X.Wu,W.Ouyang,Z.Wang,B.Ding,Y.Zhang,andG.Long,et al.,Transcription Factors BARX1 and DLX4 Contribute to ProgressionofClear Cell Renal Cell Carcinoma via Promoting Proliferation and Epithelial-Mesenchymal Transition,Front.Mol.Biosci.,vol.8,p.626328,2021.

[0150]

[10] Y.Pu,J.Han,M.Zhang,M.Liu,G.Abdusamat,and H.Liu,SKA1 promotestumor metastasis via SAFB-mediated transcription repression ofDUSP6 in clearcell renal cell carcinoma,Aging(Albany NY),vol.14,no.23,pp.9679-9698,2022.

[0151]

[11] B.Yu,X.Zheng,Z.Sun,P.Cao,J.Zhang,and W.Wang,IFI16 Can Be Used asa Biomarker for Diagnosis of Renal Cell Carcinoma and Prediction ofPatientSurvival,Front.Genet.,vol.12,p.599952,2021.

[0152]

[12] X.D.Liu,D.W.Yao,and F.Xin,TTK contributes to tumor growth andmetastasis of clear cell renal cell carcinoma by inducing cell proliferationand invasion,Neoplasma,vol.66,no.6,pp.946-953,2019.

[0153]

[13] Y.Wang,S.Liu,Y.Chen,B.Zhu,and Q.Xing,Survival Prognosis,TumorImmune Landscape,and Immune Responses of PPP1R18 in Kidney Renal Clear CellCarcinoma and Its Potentially Double Mechanisms,World J.Oncol.,vol.13,no.1,pp.27-37,2022.

[0154]

[14] H.Sun,J.Zheng,J.Xiao,J.Yue,Z.Shi,Z.Xuan,C.Chen,Y.Zhao,W.Tang,andS.Ye,et al.,TOPK / PBK is phosphorylated by ERK2 at serine 32,promotestumorigenesis and is involved in sorafenib resistance in RCC,Cell Death Dis.,vol.13,no.5,p.450,2022.

[0155]

[15] H.Fang,Z.Peng,B.Tan,N.Peng,B.Li,D.He,M.Xu,and Z.Yang,Theinvolvement ofPDIA2 gene in the progression of renal cell carcinoma ispotentially through regulation ofJNK signaling pathway,Clin.Transl.Oncol.,vol.25,no.10,pp.2938-2949,2023.

[0156]

[16] Y.Liu,H.Wang,B.Ni,J.Zhang,S.Li,Y.Huang,Y.Cai,H.Mei,and Z.Li,Lossof KCNJ15expression promotes malignant phenotypes and correlates with poorprognosis in renal carcinoma,Cancer Manag.Res.,vol.11,pp.1211-1220,2019.

[0157]

[17] S.Chen,M.Yu,L.Ju,G.Wang,K.Qian,Y.Xiao,and X.Wang,The immune-related biomarker TEK inhibits the development of clear cell renal cellcarcinoma(ccRCC)by regulating AKT phosphorylation,Cancer Cell Int.,vol.21,no.1,p.119,2021.

[0158]

[18] Y.Zhu,L.Xu,J.Zhang,W.Xu,Y.Liu,H.Yin,T.Lv,H.An,L.Liu,and H.He,etal.,Klotho suppresses tumor progression via inhibiting PI3K / Akt / GSK3β / Snailsignaling in renal cell carcinoma,Cancer Sci.,vol.104,no.6,pp.663-671,2013.

[0159]

[19] P.Xi,Z.Zhang,Y.Liu,Y.Nie,B.Gong,J.Liu,H.Huang,Z.Liu,T.Sun,andW.Xie,Multidimensional comprehensive and integrated analysis of the potentialfunction of TMEM25 in renal clear cell carcinoma with low expression status,Aging(Albany NY),vol.16,no.1,pp.367-388,2024.

[0160]

[20] J.S.Yuan,Z.S.Chen,K.Wang,and Z.L.Zhang,Holliday junction-recognition protein modulates apoptosis,cell cycle arrest and reactive oxygenspecies stress in human renal cell carcinoma,Oncol.Rep.,vol.44,no.3,pp.1246-1254,2020.

[0161]

[21] F.Zhang,D.Yuan,J.Song,W.Chen,W.Wang,G.Zhu,B.Hu,X.Chen,and J.Zhu,HJURP is a prognostic biomarker for clear cell renal cell carcinoma and islinked to immune infiltration,Int.Immunopharmacol.,vol.99,p.107899,2021.

[0162]

[22] W.Wei,Y.Lv,Z.Gan,Y.Zhang,X.Han,and Z.Xu,Identification of keygenes involved in the metastasis ofclear cell renal cell carcinoma,Oncol.Lett.,vol.17,no.5,pp.4321-4328,2019.

[0163]

[23] L.Peng,J.Liang,Q.Wang,and G.Chen,A DNA Damage Repair GeneSignature Associated With Immunotherapy Response and Clinical Prognosis inClear Cell Renal Cell Carcinoma,Front.Genet.,vol.13,p.798846,2022.

[0164]

[24] W.Zhong,C.Huang,J.Lin,M.Zhu,H.Zhong,M.H.Chiang,H.S.Chiang,M.S.Hui,Y.Lin,and J.Huang,Development and Validation ofNine-RNA BindingProtein Signature Predicting Overall Survival for Kidney Renal Clear CellCarcinoma,Front.Genet.,vol.11,p.568192,2020.

[0165]

[25] D.Pei,C.Xu,D.Wang,X.Shi,Y.Zhang,Y.Liu,J.Guo,N.Liu,and H.Zhu,ANovel Prognostic Signature Associated With the Tumor Microenvironment inKidney Renal Clear Cell Carcinoma,Front.Oncol.,vol.12,p.912155,2022.

[0166]

[26] Z.Zhang,E.Lin,H.Zhuang,L.Xie,X.Feng,J.Liu,and Y.Yu,Constructionof a novel gene-based model for prognosis prediction ofclear cell renal cellcarcinoma,Cancer Cell Int.,vol.20,p.27,2020.

[0167]

[27] Y.Yu,W.Yao,T.Wang,W.Xue,Y.Meng,L.Cai,W.Jian,Y.Yu,and C.Zhang,FBXL6 depletion restrains clear cell renal cell carcinoma progression,Transl.Oncol.,vol.26,p.101550,2022.

[0168]

[28] H.Gou,P.Chen,and W.Wu,FAM72 family proteins as poor prognosticmarkers in clear cell renal carcinoma,Biochem.Biophys.Rep.,vol.35,p.101506,2023.

[0169]

[29] B.Wang,D.Chen,and H.Hua,TBC1D3 family is a prognostic biomarkerand correlates with immune infiltration in kidney renal clear cell carcinoma,Mol.Ther.Oncolytics,vol.22,pp.528-538,2021.

[0170]

[30] Y.Zhan,R.Zhang,C.Li,X.Xu,K.Zhu,Z.Yang,J.Zheng,and Y.Guo,AmicroRNA-clinical prognosis model to predict the overall survival for kidneyrenal clear cell carcinoma,Cancer Med.,vol.10,no.17,pp.6128-6139,2021.

[0171]

[31] V.Petrozza,A.Carbone,T.Bellissimo,N.Porta,G.Palleschi,A.L.Pastore,A.Di Carlo,C.Della Rocca,and F.Fazi,Oncogenic MicroRNAsCharacterization in Clear Cell Renal Cell Carcinoma,Int.J.Mol.Sci.,vol.16,no.12,pp.29219-29225,2015.

[0172]

[32] H.Tusong,N.Maolakuerban,J.Guan,M.Rexiati,W.G.Wang,B.Azhati,Y.Nuerrula,and Y.J.Wang,Functional analysis of serum microRNAs miR-21 andmiR-106a in renal cell carcinoma,Cancer Biomark.,vol.18,no.1,pp.79-85,2017.

[0173]

[33] Z.F.Jing,J.B.Bi,Z.Li,X.Liu,J.Li,Y.Zhu,X.T.Zhang,Z.Zhang,Z.Li,andC.Z.Kong,Inhibition of miR-34a-5p can rescue disruption of the p53-DAPK axisto suppress progression of clear cell renal cell carcinoma,Mol.Oncol.,vol.13,no.10,pp.2079-2097,2019.

[0174]

[34] Y.Liu,H.Nie,Y.Zhang,N.Zhang,M.Han,H.Liu,D.Sun,X.Wu,X.Xiao,andX.Cao,MiR-224-5p Targeting OCLN Promotes the Proliferation,Migration,andInvasion of Clear Cell Renal Cell Carcinoma Cells,Urol.Int.,vol.106,no.11,pp.1185-1194,2022.

[0175]

[35] J.Zhou,P.Li,J.Feng,Q.Wu,and S.You,MiR-24-1-5p Hinders MalignantPhenotypes of Clear Cell Renal Cell Carcinoma by Targeting SHOX2,Biochem.Genet.,vol.61,no.5,pp.2004-2019,2023.

[0176]

[36] Q.Liu,and C.Lei,LINC01232 serves as a novel biomarker andpromotes tumour progression by sponging miR-204-5p and upregulating RAB22A inclear cell renal cell carcinoma,Ann.Med.,vol.53,no.1,pp.2153-2164,2021.

[0177]

[37] L.Guan,J.Tan,H.Li,and X.Jin,Biomarker identification in clearcell renal cell carcinoma based on miRNA-seq and digital gene expression-seqdata,Gene,vol.647,pp.205-212,2018.

[0178]

[38] S.Qin,X.Shi,C.Wang,P.Jin,and F.Ma,Transcription Factor and miRNAInterplays Can Manifest the Survival ofccRCC Patients,Cancers(Basel),vol.11,no.11,p.1668,2019.

[0179]

[39] P.Hong,H.Du,M.Tong,Q.Cao,D.Hu,J.Ma,Y.Jin,Z.Li,W.Huang,and G.Tong,A Novel M7G-Related MicroRNAs Risk Signature Predicts the Prognosis and TumorMicroenvironment of Kidney Renal Clear Cell Carcinoma,Front.Genet.,vol.13,p.922358,2022.

[0180]

[40] M.Huang,T.Zhang,Z.Y.Yao,C.Xing,Q.Wu,Y.W.Liu,and X.L.Xing,MicroRNArelated prognosis biomarkers from high throughput sequencing data ofkidneyrenal clear cell carcinoma,BMC Med.Genomics,vol.14,no.1,p.72,2021.

[0181]

[41] M.Han,H.Yan,K.Yang,B.Fan,P.Liu,and H.Yang,Identificationofbiomarkers and construction of a microRNA-mRNA regulatory network for clearcell renal cell carcinoma using integrated bioinformatics analysis,PLoS One,vol.16,no.1,p.e0244394,2021.

[0182]

[42] X.Zuo,C.Lu,Y.Zheng,D.Lai,D.Liu,G.Wan,C.Lu,and X.Gu,Effects oftheTargeted Regulation of CCRK by miR-335-5p on the Proliferation andTumorigenicity of Human Renal Carcinoma Cells,J.Oncol.,vol.2022,p.2960050,2022.

Claims

1. A renal cell carcinoma biomarker, characterized in that The markers are one or more of the following: Among them, Δ represents the up-regulation of DEGs-prognostic or DEMs-prognostic predicted in this study, Represents the down-regulation of DEGs-prognostic or DEMs-prognostic predicted in this study.

2. A prognosis gene for renal cell carcinoma, characterized in that: The gene is one or more of 17 prognostic genes, and the 17 prognostic genes include 14 upregulated prognostic genes: CCDC194, BORCS8-MEF2B, CST2, CDKN2C, ZIC5, PCDHGC5, ERICH6B, ZACN, MXRA8, RPP21, COL6A2, DEPP1, C1orf162, PLEKHG4 and 3 downregulated prognostic genes: TMEM200C, FOXE3, IL13RA2.

3. A method for screening biomarkers for renal cell carcinoma, characterized in that: The screening method is to screen through a renal cell carcinoma prognostic model, wherein the screening specifically constructs a stratified prognostic model based on a characteristic system of race, pathological stage, and whether lymph nodes are metastatic, and then evaluates the prognostic model to obtain renal cell carcinoma biomarkers included in the model; The method for constructing the renal cell carcinoma prognosis model comprises the following steps: (1) Obtain renal cell carcinoma-related mRNA and miRNA, group them, analyze them, and optimize them; (2) Prognostic models were established by multivariate Cox regression analysis, including mRNA, miRNA, or mRNA+miRNA prognostic models; (3) Evaluate the prognostic model.

4. The method according to claim 3, characterized in that: The grouping in the construction method is to group the samples according to the characteristics of race, pathological stage and whether lymph node metastasis occurs; the grouping includes total RCC samples, white samples, black samples, Stage I-II samples, Stage III-IV samples, white stage I-II samples, white stage III-IV samples, black stage I-II samples, black stage III-IV samples, samples without lymph node metastasis, lymph node metastasis samples, white samples without lymph node metastasis, white samples with lymph node metastasis and black samples with lymph node metastasis.

5. The method according to claim 3, characterized in that: The analysis in the construction method is: using the R package "edgeR" to perform differential expression analysis, screen out differentially expressed mRNAs and miRNAs, and name them as DEGs and DEMs, respectively.

6. The method according to claim 5, characterized in that: The analysis also includes: performing univariate Cox analysis and KM analysis on DEGs and DEMs, and taking the intersection of the two sets of analysis results to obtain DEGs-os and DEMs-os, respectively.

7. The method according to claim 3, characterized in that: The optimization in the construction method is: optimizing DEGs-os and DEMs-os respectively through Lasso analysis, and merging the optimized DEGs-os and DEMs-os matrices to obtain DEGs-os+DEMs-os.

8. The method according to claim 3, characterized in that: The method for establishing the prognostic model was as follows: multivariate Cox proportional hazard regression analysis was performed on DEGs-os, DEMs-os and DEGs-os+DEMs-os markers using the "survminer" package and a prognostic model was constructed.

9. The method according to claim 3, characterized in that: Evaluation of the prognostic model includes: dividing RCC patients into high-risk and low-risk groups by median risk score, drawing risk score curves and expression heat maps, and drawing survival curves of patients in high-risk and low-risk groups by KM analysis; or, drawing ROC curves of mRNA, miRNA and mRNA+miRNA prognostic models to evaluate their predictive ability; or, using Spearman correlation analysis to evaluate the relationship between the stratification depth of the RCC prognostic model and the three-year and five-year survival AUC values.

10. The method according to claim 3, characterized in that: The method for constructing the prognostic model also includes: using literature search to check the prognostic mRNA and miRNA obtained, and comparing their expression with the corresponding prognostic RNA in existing RCC-related literature.