Hepatocellular carcinoma prognostic marker based on innate immune gene and screening method of hepatocellular carcinoma prognostic marker

Abstract

The invention belongs to the technical field of biomedicine, and discloses a hepatocellular carcinoma prognostic marker based on innate immune genes and a screening method thereof, a group of innate immune genes (16 genes) closely related to hepatocellular carcinoma is screened and quantified as the prognostic marker, the five-year survival rate of hepatocellular carcinoma patients is effectively predicted, and the prognostic marker has the advantages that the prognostic marker is high in specificity and high in sensitivity. Particularly, the method has a remarkable effect on predicting the survival rates of one year, two years and three years, and the effect of the method is better than that of an existing clinical marker alpha fetoprotein (AFP).

Description

Technical Field

[0001] The invention belongs to the field of biomedical technology, and particularly relates to a group of innate immune genes closely related to hepatocellular carcinoma and quantifying them as prognostic markers, as well as a method for screening the prognostic markers. Background Art

[0002] The mortality rate of hepatocellular carcinoma (HCC) ranks third in the world and second in my country, with a five-year survival rate of approximately 18% [1,2]. Currently, immunotherapy has become a breakthrough in the treatment of hepatocellular carcinoma, but the number of patients who benefit from it is less than 30% [3]. The reason is that too much attention has been paid to the single factor of adaptive immune response. As research deepens, it has been found that tumor immunity is a dynamic balance involving many innate immune factors [4]. Effective innate immunity can directly induce innate immune cytotoxic effects and indirectly induce adaptive immune cytotoxic effects to kill tumor cells [5], thereby increasing the sensitivity of immunotherapy and effectively improving prognosis [6]. AFP is a common biomarker for hepatocellular carcinoma, but its sensitivity and specificity are relatively low.

[0003] After searching, the Chinese invention patent (CN115019965A) disclosed a method for constructing a survival prediction model for hepatocellular carcinoma patients based on cell death-related genes. This method can accurately predict the survival of hepatocellular carcinoma patients based on the prognostic model of three new programmed cell death-related genes, providing a new direction and strategy for the grouping and stratification management and precision treatment of hepatocellular carcinoma; the risk index Risk Index can be used as an independent prognostic factor to predict the survival of hepatocellular carcinoma patients, and the universal applicability of the model was externally verified through another public database ICGC database as a validation set, providing a new direction for the diagnosis and treatment of hepatocellular carcinoma. This patent selects cell death-related genes and constructs a prediction model using univariate Cox regression and multivariate Cox regression analysis.

[0004] However, there are no reports on the screening of key genes from innate immune genes to predict the survival of hepatocellular carcinoma. The present invention intends to use a group of innate immune genes to predict survival, and is expected to guide the clinical treatment of hepatocellular carcinoma, and increase the number of beneficiaries of various treatment options including immunotherapy by reshaping the microenvironment.

[0005] During further research, we found that the results were not ideal when directly using the GEO (gene expression data) dataset as external validation. The reason was that patients in different datasets may come from different regions or hospitals, patients may have different disease subtypes or treatments, different sequencing platforms and methods, etc. These factors cause their data to be different from the data benchmark of the training set, making the external validation effect inaccurate.

[0006] References [1] Bray F, Laversanne M, Sung H, Ferlay J, Siegel RL, Soerjomataram I, Jemal A. Global cancer statistics 2022: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin. 2024 May-Jun;74(3):229-263. [2] Hao X, Sun G, Zhang Y, Kong X, Rong D, Song J, Tang W, Wang X. Targeting Immune Cells in the Tumor Microenvironment of Hepatocellular Carcinoma: New Opportunities and Challenges. Front Cell Dev Biol. 2021 Nov 12;9:775462. doi:10.3389 / fcell.2021.775462. PMID: 34869376; PMCID: PMC8633569. [3] Mishra DK, Rocha HJ, Miller R, Kim MP. Immune cells inhibit the tumor metastasis in the 4D cellular lung model by reducing the number of live circulating tumor cells. Sci Rep. 2018 Nov 8;8(1):16569. doi: 10.1038 / s41598-018-34983-7. PMID: 30410108; PMCID: PMC6224572. [4] Chen DS, Mellman I. Elements of cancer immunity and the cancer-immune set point. Nature. 2017 Jan 18;541(7637):321-330. doi: 10.1038 / nature21349. PMID: 28102259. [5] Maiorino L, Daßler-Plenker J, Sun L, Egeblad M. Innate Immunityand Cancer Pathophysiology. Annu Rev Pathol. 2022 Jan 24;17:425-457. doi:10.1146 / annurev-pathmechdis-032221-115501. Epub 2021 Nov 17. PMID: 34788549;PMCID: PMC9012188. [6] Lopes JA, Borges-Canha M, Pimentel-Nunes P. Innate immunity andhepatocarcinoma: Can toll-like receptors open the door to oncogenesis? WorldJ Hepatol. 2016 Jan 28;8(3):162-82. doi: 10.4254 / wjh.v8.i3.162. PMID:26839640; PMCID: PMC4724579. Summary of the invention

[0007] The object of the present invention is to provide a prognostic marker for hepatocellular carcinoma based on innate immune genes, including a group of innate immune-related genes, namely AOC1, SOCS2, S100A9, TNFRSF11B, IL18RAP, BLNK, RAP1GAP2, PSMD4, IL12A, C1orf35, MMP1, IRAK1, CCL20, KPNA2, LPCAT1, and TNFRSF4.

[0008] Preferably, the risk score is calculated based on the expression level and risk coefficient of the group of innate immunity-related genes, as follows: Risk score = 0.27257418 × AOC1 gene expression level - 0.36883064 × SOCS2 gene expression level + 0.19643025 × S100A9 gene expression level + 0.02618757 × TNFRSF11B gene expression level - 1.33919446 × IL18RAP gene expression level - 0.23234786 × BLNK gene expression level + 0.15603121 × RAP1GAP2 gene expression level + 0.18901848 × PSMD4 gene expression level + 0.38220082 × IL12A gene expression level + 0.07473103 × C1orf35 gene expression level + 0.23048984 × MMP1 gene expression level + 0.39702235 × IRAK1 gene expression level + 0.00363004×CCL20 gene expression level + 0.00564114×KPNA2 gene expression level + 0.49308590×LPCAT1 gene expression level + 0.22578477×TNFRSF4 gene expression level.

[0009] Another object of the present invention is to provide a method for screening prognostic markers for hepatocellular carcinoma based on innate immune genes, which is used to screen the above-mentioned prognostic markers, comprising the following steps: S1, compare hepatocellular carcinoma tissues and normal tissues in the TCGA database to obtain differentially expressed genes in hepatocellular carcinoma tissues, search for innate immune genes on the Gene Cards website, organize them into gene sets, and take the intersection of hepatocellular carcinoma differentially expressed genes and innate immune genes; S2, TCGA and GEO were used as model construction and external validation databases, and their data were normalized and batch corrected; S3, univariate Cox regression analysis was performed on the gene expression data and clinical data of HCC patients obtained from the TCGA database to identify genes in the innate immunity-related gene set that were associated with patient survival; S4. The patients of TCGA-LIHC were randomly divided into a training set and an internal validation set in a ratio of 7:3. The genes related to survival obtained in the previous step were further processed by Lasso regression analysis. By introducing L1 regularization, variable selection was automatically performed, and the most relevant part was screened out from many genes to obtain the prognostic markers of hepatocellular carcinoma based on innate immune genes. The prognostic markers included the group of innate immune-related genes, and their corresponding risk coefficients, and the risk scores calculated using the expression levels of the innate immune-related genes and the corresponding risk coefficients.

[0010] Preferably, the expression levels and risk coefficients of relevant genes in the TCGA database are calculated to obtain the risk score for each patient, and the risk score is used to analyze the patient's survival: the hepatocellular carcinoma patient dataset GSE14520 in the GEO database is used as an external validation to verify the survival prediction effect of the model and draw a Kaplan-Meier curve.

[0011] Preferably, for the risk score and the patient's clinical data, a time-dependent ROC curve is used to evaluate the predictive performance of the risk score at different time points, and a risk score graph is plotted to visualize the predictive performance.

[0012] Preferably, GO analysis is performed on the group of innate immunity-related genes to reveal that the activities of NF-κB transcription factors and natural killer cell-mediated immune effects are the main factors affecting the survival of patients with hepatocellular carcinoma.

[0013] Preferably, in step S2, first, the TCGA and GEO datasets are corrected for batch effects using the ComBat method to eliminate technical differences between experimental platforms or batches, and then the data are standardized using normalize Between Arrays to ensure comparability between different datasets.

[0014] The present invention has the following beneficial effects: The present invention is based on the hepatocellular carcinoma prognostic marker and screening method of innate immune genes, and screens out a group of innate immune genes (16 genes) closely related to hepatocellular carcinoma as prognostic markers for predicting survival and guiding clinical treatment. Its performance is stronger than the existing clinical marker AFP (alpha fetoprotein). In the later stage, it is expected to increase the number of beneficiaries of various treatment options including immunotherapy by reshaping innate immunity.

[0015] The present invention uses R language as a tool to perform univariate Cox regression analysis, Lasso regression analysis, patient risk stratification, Kaplan-Meier survival curve drawing, ROC curve drawing, etc., to screen and obtain genes closely related to survival, quantify them, and analyze their main functions. The prognostic markers screened by the present invention have a good predictive effect on the one-year, two-year, and three-year survival of patients with hepatocellular carcinoma, verifying the important role of the activity of NF-κB transcription factor and the immune effect mediated by natural killer cells in the prognosis of hepatocellular carcinoma; the combination of Lasso based on univariate Cox regression is suitable for processing high-dimensional data, automatically selecting variables, and can effectively solve the problem of multicollinearity, while reducing overfitting and enhancing the generalization ability of the model.

[0016] We used TCGA (liver cancer data) and GEO (gene expression data) as the main data sources. First, we corrected the batch effects of these two datasets by the ComBat method to eliminate technical differences between experimental platforms or batches. Subsequently, the data were standardized by normalize Between Arrays to ensure comparability between different datasets. We used univariate Cox regression analysis to screen out genes related to survival, and further optimized feature selection by the Lasso regression model, which ultimately improved the predictive performance of the model. External validation showed that batch effect correction and standardization significantly enhanced the generalization ability of the model. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 The present invention is a flowchart of the method for screening hepatocellular carcinoma prognostic markers based on innate immune genes.

[0018] Figure 2 4 is a flow chart of the method for screening prognostic markers in an embodiment of the present invention.

[0019] Figure 3 It is the volcano map of differentially expressed genes between hepatocellular carcinoma tissue and normal tissue in the present invention.

[0020] Figure 4 It is a Venn diagram of the innate immunity-related genes and the differentially expressed genes in hepatocellular carcinoma in the present invention.

[0021] Figure 5 and Figure 6 This is a diagram showing the effect of batch correction of data from the TCGA and GEO databases in the present invention.

[0022] Figure 7 This is a result diagram of screening survival-related genes by single-factor Cox regression of the genes obtained from the Venn diagram in the present invention.

[0023] Figure 8 and Fig. 9 This is a process diagram of further screening key genes using Lasso regression from 112 genes screened by single factor Cox regression in the present invention.

[0024] Fig.10 It is the Kaplan-Meier curve diagram of the training set in the present invention.

[0025] Fig.11 It is the Kaplan-Meier curve diagram of the internal validation set in the present invention.

[0026] Fig.12 It is the Kaplan-Meier curve diagram of the external validation set in the present invention.

[0027] Fig.13It is the risk score graph of the training set in the present invention.

[0028] Fig.14 It is the risk score graph of the internal validation set in the present invention.

[0029] Fig.15 It is the risk score graph of the external validation set in the present invention.

[0030] Fig.16 It is the ROC curve diagram of the training set in the present invention.

[0031] Fig.17 It is the ROC curve diagram of the internal validation set in the present invention.

[0032] Fig.18 It is the ROC curve diagram of the external validation set in the present invention.

[0033] Fig.19 It is the ROC curve diagram of AFP in the TCGA database in the present invention.

[0034] Fig. 20 GO analysis of 16 key genes in the present invention DETAILED DESCRIPTION

[0035] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0036] like Figure 1 , Figure 2 As shown, a method for screening prognostic markers for hepatocellular carcinoma based on innate immune genes, the steps are as follows: S1: Using the hepatocellular carcinoma RNA data in the public database TCGA database, such as Figure 3 The differential gene analysis based on Counts data of hepatocellular carcinoma and normal tissues was shown, and a total of 5151 differentially expressed genes were obtained. From the Gene Cards website, innate immune-related genes were searched, and a total of 2025 genes were obtained. Figure 4 As shown, the intersection of innate immunity-related genes and differentially expressed genes resulted in a total of 370 innate immunity-related genes differentially expressed in hepatocellular carcinoma tissues.

[0037] S2: The GSE14520 dataset from the GEO database was selected as an external validation. Before subsequent analysis, the TCGA and GEO data needed to be normalized and batch corrected. The TCGA database used the FPKM method for normalization, and the data downloaded from the GEO database had been processed by RMA. Then, the same genes were taken from the data of the two databases for batch correction. The effect of batch correction is shown in Figure 2. Figure 5 , Figure 6 shown.

[0038] TCGA (liver cancer data) and GEO (gene expression data) were used as the main data sources. First, the two datasets were corrected for batch effects using the ComBat method to eliminate technical differences between experimental platforms or batches. Subsequently, the data were standardized using normalize Between Arrays to ensure comparability between different datasets.

[0039] S3: Figure 7 As shown, we performed univariate Cox regression analysis on the above 370 differentially expressed innate immunity-related genes, and combined the batch-corrected transcriptome data with the clinical data of hepatocellular carcinoma patients obtained from the TCGA database for univariate Cox regression analysis to determine the genes associated with overall survival; after this step, 112 innate immunity-related differential genes associated with the prognosis of hepatocellular carcinoma were obtained.

[0040] S4: Figure 8 , Fig. 9 As shown, the 112 genes screened by univariate Cox regression were further screened by Lasso regression, and the patients of TCGA-LIHC were randomly divided into training set and internal validation set in a ratio of 7:3. The Lasso-Cox regression model was fitted using the training set data, the optimal lambda value was selected, and the coefficients of the corresponding model were extracted. Finally, the prognostic markers were obtained, including 16 innate immunity-related genes and their corresponding risk coefficients.

[0041] Univariate Cox regression analysis was used to screen out genes associated with survival, and Lasso regression was used to further optimize gene selection, ultimately improving the predictive effect of the risk score. External validation showed that batch effect correction and standardization made the data benchmarks of different databases uniform and comparable.

[0042] S5: The risk score was calculated based on the model obtained by Lasso regression, and the risk score was used as another prognostic marker. The model was validated on the internal validation set, the C-index was calculated, and risk stratification analysis was performed. The Kaplan-Meier curve and risk score graph were drawn to visualize the survival differences of different risk groups. Fig.10 , Fig.11 As shown in, they are the Kaplan-Meier curves of the training set and the internal validation set, respectively. Fig.13 , 14As shown in the figure, they are the risk score graphs of the training set and the internal validation set, respectively. According to the results of Lasso regression analysis, the GSE14520 dataset related to hepatocellular carcinoma in the GEO database was used as the external validation set. The risk score (Risk Score) was calculated based on the gene expression results and risk correlation coefficients of the above batch corrections. The same processing was performed as the training set and the internal validation set, and the Kaplan-Meier curve and risk score graph were drawn to visualize the survival differences of different risk groups, as shown in the figure. Fig.12 and 15 Shown are the Kaplan-Meier curves and risk score plots of the external validation set, respectively.

[0043] S6: For the risk scores and the patients’ clinical data, we used time-dependent ROC curves to plot the data from the two databases at five time points: one year, two years, three years, four years, and five years, to evaluate the predictive effect of the risk scores at different time points, such as Figure 16-18 As shown, and compared with AFP, such as Fig.19 shown.

[0044] S7: We performed GO analysis on the 16 genes obtained from Lasso regression and found that Fig. 20 As shown, in terms of biological processes analyzed by GO, the most relevant ones are the activity of NF-κB transcription factor and natural killer cell-mediated immune effects.

[0045] The NF-κB signaling pathway has a dual role in the innate immunity of hepatocellular carcinoma. On the one hand, it exerts anti-tumor effects by regulating inflammation and innate immune cell functions; on the other hand, it is used in the hepatocellular carcinoma microenvironment to promote immune escape and tumor progression. And its specific role is related to the degree of activation of the NF-κB signaling pathway. Excessive activation of NF-κB signaling not only promotes tumor cell survival, invasion, metastasis, genomic instability and metabolic abnormalities, but also reshapes the immunosuppressive microenvironment, promotes immune escape and resistance to immunotherapy. Therefore, regulatory strategies targeting the NF-κB signaling pathway may provide a new direction for the treatment of hepatocellular carcinoma.

[0046] Natural killer (NK) cells play an important role in the immune surveillance of liver cancer, inhibiting tumor growth by recognizing and killing tumor cells and secreting cytokines (such as interferon-γ). However, the immune microenvironment of liver cancer often inhibits the function of NK cells through immunosuppressive factors, leading to immune escape and tumor progression. The number and functional status of NK cells are closely related to the clinical prognosis of liver cancer. Stronger NK cell activity is usually associated with a better prognosis. Therefore, it is of great significance as a potential prognostic marker in immunotherapy of liver cancer.

[0047] In a specific example of the present invention, the 16 innate immune-related genes after screening are AOC1, SOCS2, S100A9, TNFRSF11B, IL18RAP, BLNK, RAP1GAP2, PSMD4, IL12A, C1orf35, MMP1, IRAK1, CCL20, KPNA2, LPCAT1, and TNFRSF4; the relevant expression levels of the above 16 genes and the risk correlation coefficient are calculated to obtain a risk score, and the calculation formula is: Risk Score = 0.27257418 × AOC1 gene expression level - 0.36883064 × SOCS2 gene expression level + 0.19643025 × S100A9 gene expression level + 0.02618757 × TNFRSF11B gene expression level - 1.33919446 × IL18RAP gene expression level - 0.23234786 × BLNK gene expression level + 0.15603121×RAP1GAP2 gene expression level+ 0.18901848×PSMD4 gene expression level+0.38220082×IL12A gene expression level+0.07473103×C1orf35 gene expression level+ 0.23048984×MMP1 gene expression level+ 0.39702235×IRAK1 gene expression level+ 0.00363004×CCL20 gene expression level+ 0.00564114×KPNA2 gene expression level+ 0.49308590×LPCAT1 gene expression level+ 0.22578477×TNFRSF4 gene expression level.

[0048] Among the 16 key prognostic genes for hepatocellular carcinoma screened, AOC1, BLNK, RAP1GAP2, and C1orf35 have not yet been clearly identified as prognostic markers for hepatocellular carcinoma, and there are few studies on them in hepatocellular carcinoma. However, our research results found that they are differentially expressed in hepatocellular carcinoma and have a good predictive effect on the survival of hepatocellular carcinoma patients.

[0049] In a specific example of the present invention, the patients in the TCGA cohort were randomly divided into a training set and an internal validation set at a ratio of 7:3, wherein the training set had 153 patients and the internal validation set had 109 patients. The training set and the internal validation set were divided into a high-risk group and a low-risk group based on the median risk score value. The risk index of each gene obtained by Lasso regression analysis was combined with the risk score prediction model, and the prediction effect of the model was visualized to draw the Kaplan-Meier survival curve and the risk score graph; the GSE14520 data set of the GEO database was used as an external validation, and a total of 221 hepatocellular carcinoma patients with survival information in the screening data set were screened. Similarly, the risk index of each gene obtained by Lasso regression analysis was combined with the risk score prediction model, and the prediction effect of the model was visualized to draw the Kaplan-Meier survival curve and the risk score graph. At the same time, the TCGA and GEO databases were evaluated by the time-dependent ROC curve to evaluate the predictive performance of the risk score for the total survival of the survival time. The time-dependent ROC analysis and the timeROC and survival packages of the survival analysis were used to draw the ROC curves of one, two, three, four, and five years for the two databases.

[0050] In the specific example of the present invention, the Kaplan–Meier survival curve diagram is divided into two groups according to the risk scores corresponding to the patients in each data set, with the median risk score as the boundary, which are defined as High and Low respectively. From the figure, we can clearly observe that there is a significant difference in the survival of the two groups whether in the training set, the internal validation set or the external validation set.

[0051] In the specific example of the present invention, the risk score graph A is arranged from low to high according to the risk score corresponding to each patient, and the corresponding survival status of each patient is shown in the graph. It can be observed from the graph that in the training set, the internal validation set, and the external validation set, the patients with high risk scores have a shorter overall survival than the patients with low risk scores. The corresponding gene expression heat map constituting the survival model is shown in the graph C. In the specific example of the present invention, time-dependent ROC curves are drawn for the TCGA and GEO databases, respectively, to test the effect of risk scores on the one-year, two-year, three-year, four-year, and five-year survival prediction of hepatocellular carcinoma patients in the two databases. Fig.16 It can be observed that the area under the curve (AUC) of the training set reaches 0.84, 0.82, 0.82, 0.86, and 0.90 in one, two, three, four, and five years, respectively. Fig.17 It can be observed that the area under the curve (AUC) of the internal validation set reaches 0.77, 0.82, 0.76, 0.67, and 0.61 in one, two, three, four, and five years, respectively. Fig.18It can be observed that the area under the curve (AUC) of the external validation set reaches 0.63, 0.69, 0.67, 0.66, and 0.64 in one, two, three, four, and five years, respectively. Fig.19 It can be observed that the area under the curve (AUC) of AFP in the TCGA database reached 0.62, 0.53, 0.54, 0.55, and 0.51 at one, two, three, four, and five years, respectively. The risk score has a better predictive effect on the survival of hepatocellular carcinoma patients than AFP.

[0052] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A prognostic marker for hepatocellular carcinoma based on innate immune genes, characterized in that: It includes a group of innate immunity-related genes, namely AOC1, SOCS2, S100A9, TNFRSF11B, IL18RAP, BLNK, RAP1GAP2, PSMD4, IL12A, C1orf35, MMP1, IRAK1, CCL20, KPNA2, LPCAT1, and TNFRSF4.

2. The hepatocellular carcinoma prognostic marker based on innate immune genes according to claim 1, characterized in that: The risk score (Risk Score) is calculated according to the expression level and risk coefficient of the set of innate immunity-related genes, as follows: Risk score = 0.27257418 × AOC1 gene expression level - 0.36883064 × SOCS2 gene expression level + 0.19643025 × S100A9 gene expression level + 0.02618757 × TNFRSF11B gene expression level - 1.33919446 × IL18RAP gene expression level - 0.23234786 × BLNK gene expression level + 0.15603121 × RAP1GAP2 gene expression level + 0.18901848 × PSMD4 gene expression level + 0.38220082 × IL12A gene expression level + 0.07473103 × C1orf35 gene expression level + 0.23048984 × MMP1 gene expression level + 0.39702235 × IRAK1 gene expression level + 0.00363004×CCL20 gene expression level + 0.00564114×KPNA2 gene expression level + 0.49308590×LPCAT1 gene expression level + 0.22578477×TNFRSF4 gene expression level.

3. A method for screening hepatocellular carcinoma prognostic markers based on innate immune genes, used for screening the prognostic markers according to claim 1 or 2, characterized in that: The following steps are involved: S1, compare hepatocellular carcinoma tissues and normal tissues in the TCGA database to obtain differentially expressed genes in hepatocellular carcinoma tissues, search for innate immune genes on the Gene Cards website, organize them into gene sets, and take the intersection of hepatocellular carcinoma differentially expressed genes and innate immune genes; S2, TCGA and GEO were used as model construction and external validation databases, and their data were normalized and batch corrected; S3, univariate Cox regression analysis was performed on the gene expression data and clinical data of HCC patients obtained from the TCGA database to identify genes in the innate immunity-related gene set that were associated with patient survival; S4. The patients of TCGA-LIHC were randomly divided into a training set and an internal validation set in a ratio of 7:

3. The genes related to survival obtained in the previous step were further processed by Lasso regression analysis. By introducing L1 regularization, variable selection was automatically performed, and the most relevant part was screened out from many genes to obtain the prognostic markers of hepatocellular carcinoma based on innate immune genes. The prognostic markers included the group of innate immune-related genes, and their corresponding risk coefficients, and the risk scores calculated using the expression levels of the innate immune-related genes and the corresponding risk coefficients.

4. The method for screening hepatocellular carcinoma prognostic markers based on innate immune genes according to claim 3, characterized in that: The expression levels and risk coefficients of relevant genes in the TCGA database were calculated to obtain the risk score for each patient, and the risk score was used to analyze the patient's survival: the hepatocellular carcinoma patient dataset GSE14520 from the GEO database was used as an external validation to verify the survival prediction effect of the model and draw the Kaplan-Meier curve.

5. The method for screening hepatocellular carcinoma prognostic markers based on innate immune genes according to claim 3, characterized in that: For the risk score and the patient's clinical data, the time-dependent ROC curve was used to evaluate the predictive performance of the risk score at different time points, and the risk score graph was plotted to visualize the predictive performance.

6. The method for screening hepatocellular carcinoma prognostic markers based on innate immune genes according to claim 3, characterized in that: GO analysis of the group of innate immunity-related genes revealed that the activity of NF-κB transcription factor and the immune effect mediated by natural killer cells were the main factors affecting the survival of patients with hepatocellular carcinoma.

7. The method for screening hepatocellular carcinoma prognostic markers based on innate immune genes according to claim 3, characterized in that: In step S2, first, the TCGA and GEO datasets were corrected for batch effects using the ComBat method, and then the data were standardized using `normalize Between Arrays`.

Images