Hepatocellular carcinoma prognosis model construction method

Abstract

The invention discloses a hepatocellular carcinoma prognosis model construction method, and relates to the technical field of biomedicine. Comprising the following steps: S1, data acquisition and preprocessing; s2, identifying and screening m6A methylation regulatory factors; s3, performing immunohistochemistry and immunofluorescence analysis; s4, carrying out correlation analysis on the differential long-chain non-coding RNA and the LRPPRC; s5, constructing and verifying a risk model; s6, performing independent prognosis analysis and column diagram construction; and S7, performing function enrichment analysis. According to the invention, the m6A methylation regulation factor LRPPRC is combined with LncRNAs to construct a hepatocellular carcinoma prognosis model, so that a new view angle and method are provided for prognosis evaluation; through a mode of combining bioinformatics analysis and experimental verification, LncRNAs related to LRPPRC, which are highly related to survival and prognosis of hepatocellular carcinoma patients, are screened out; the constructed risk model can accurately distinguish a high-risk group and a low-risk group.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of biomedical technology, and particularly relates to a hepatocellular carcinoma prognosis model construction method. BACKGROUND

[0002] N6-methyladenosine (m6A) is a new type of reversible RNA epigenetic modification discovered in recent years, which is abundant in mRNA and non-coding RNA (ncRNA). m6A affects the fate of RNA by participating in RNA splicing, localization, stability, translation and other processes, and plays an important role in biological processes such as cancer occurrence and development. The modification of m6A depends on three regulatory factors: methyltransferase (writer), such as METTL3, WTAP, ZC3H13, KIAA1429, RBM15B, RBM15, METTL16, METTL14, etc.; demethylase (eraser), such as FTO, ALKBH5, etc.; and m6A binding protein (reader), such as YTHDF1, YTHDF2, YTHDF3, YTHDC1, YTHDC2, RBMX, IGF2BP1, IGF2BP2, IGF2BP3, HNRNPC, HNRNPA2B1, LRPPRC, FMR1, etc., which play an important role in different tumors. Two of the three regulatory factors can interact to affect tumor development. Existing studies have found that METTL3 inhibits the expression of SOCS2 through the m6A-YTHDF2 mechanism, thereby affecting tumor development. They can also affect tumor cell proliferation by regulating cell cycle conversion. Existing studies have concluded that SNHG17 interacts with LRPPRC, stabilizes c-Myc protein, promotes G1 / S phase conversion and tumor cell proliferation. In addition, they can also affect tumor drug resistance. Existing studies have shown that increased expression of METTL14 can increase the level of cytidine deaminase (CDA), thereby inactivating gemcitabine and promoting pancreatic cancer drug resistance.

[0003] Hepatocellular carcinoma (HCC) is a common tumor with a high mortality rate. Traditional surgical methods have poor efficacy, leading to poor prognosis of HCC. In recent years, the role of m6A methylation modification in HCC has attracted widespread attention. For example, in terms of the influence of m6A on the occurrence, development and metastasis of HCC, existing studies have found that METTL14 regulates primary microRNA126 through m6A modification, thereby inhibiting liver cancer metastasis. In terms of the influence of HCC on drug resistance, existing studies have shown that METTL3 deletion affects autophagy by destabilizing FOXO3 mRNA, thereby enhancing the drug resistance of HCC to sorafenib. In terms of HCC immunity, existing studies have found that the down-regulation of non-coding RNA miR-362-3p and miR-425-5p mediated ZC3H13 is associated with poor prognosis and tumor immune infiltration of hepatocellular carcinoma. More and more studies have shown that a variety of biomarkers play an important role in the diagnosis, treatment and prognosis of tumors. However, the role of m6A methylation regulatory factors as biomarkers in the diagnosis, treatment and prognosis of HCC remains to be studied.

[0004] Long non-coding RNA (LncRNA) is longer than 200 nucleotides and cannot be translated into functional proteins. It mainly affects the development of various characteristics of tumors, including proliferation, survival, and relationship with tumor microenvironment, through specific interaction with DNA, RNA and protein. For example, the interaction between lncRNA and miRNA in tumor genetics and drug resistance. Bracco et al. found that the specific regulation of DNA methylation of lncRNA MEG3 by Mir-29a may promote the growth of hepatocellular carcinoma (HCC), and the relationship between miRNA and lncRNA and the epigenetic regulation of gene expression have also been confirmed. Existing studies have shown that sufficient LINC01273 enhances the inhibition of METTL3 mRNA by increasing the stability of miR-600, leading to sorafenib resistance in hepatoma cells. In addition, some researchers use lncRNA as a diagnostic and prognostic marker for hepatocellular carcinoma. With further research, researchers have found that m6A can change the conformation of lncRNA, thereby regulating the stability and expression of lncRNA and affecting the occurrence and development of tumors. Existing studies have constructed a lncRNA model to predict the survival rate of hepatocellular carcinoma, and used the ceRNA network to point out the miRNA and mRNA related to lncRNA. Existing studies have found that lipopolysaccharide regulates PD-L1 through m6A modification of lncRNA (MIR155HG), thereby promoting the immune escape of hepatoma cells.

[0005] At present, the interaction between m6A regulatory factors and lncRNA in hepatocellular carcinoma has not been fully studied, and its ability in the diagnosis and survival and prognosis prediction of hepatocellular carcinoma is still unclear. SUMMARY

[0006] The application aims at providing a hepatocellular carcinoma prognosis model construction method to solve the problems in the prior art.

[0007] To achieve the above-mentioned purpose, the application adopts the following technical solutions: A hepatocellular carcinoma prognosis model construction method, comprising the following steps: S1: data acquisition and preprocessing; S2: identification and screening of m6A methylation regulatory factors; S3: immunohistochemical and immunofluorescence analysis; S4: differential long non-coding RNA and LRPPRC correlation analysis; S5: risk model construction and verification; S6: independent prognosis analysis and nomogram construction; S7: functional enrichment analysis.

[0008] Preferably, in S1, the data acquisition and preprocessing comprises: S12: downloading RNA sequencing data, clinical information and survival data of hepatocellular carcinoma patients and normal controls from the UCSC Xena database; S13: cleaning the data to remove invalid or missing data entries; S14: standardizing the data to ensure consistency of data from different sources; S15: processing missing values and filling missing data by using appropriate imputation methods (such as mean replacement, median replacement, etc.).

[0009] Preferably, in S2, the identification and screening of m6A methylation regulatory factors comprises: S21: determining a list of 23 m6A methylation regulatory factors; S22: extracting the expression matrix of these regulatory factors in hepatocellular carcinoma patients and normal controls; S23: using the "heatmap" package of R language to draw gene expression heat map to show the expression of regulatory factors in different samples; S24: using the "survival" package of R language to perform univariate Cox regression analysis to preliminarily screen out regulatory factors related to survival; S25: performing multivariate Cox regression analysis to screen regulatory factors with independent prognostic value.

[0010] Preferably, in S3, the immunohistochemical and immunofluorescence analysis comprises: S31: collecting tissue samples of hepatocellular carcinoma patients; S32: Sectioning the tissue sample and performing fixation and dehydration; S33: Performing immunohistochemical staining, using anti-LRPPRC antibody to label tumor cells, observing and recording the expression pattern of LRPPRC under microscope; S34: Performing immunofluorescence analysis, using anti-LRPPRC antibody combined with fluorescent markers, observing and recording the expression of LRPPRC under fluorescence microscope.

[0011] Preferably, in S4, the LRPPRC correlation analysis of differential long non-coding RNAs includes: S41: Obtain the annotation file of all LncRNAs from GENCODE database; S42: Perform differential expression analysis using the "DESeq2" package of R language to identify significantly differentially expressed LncRNAs in hepatocellular carcinoma; S43: Calculate the Pearson correlation coefficient between the differential LncRNAs and LRPPRC using the "corrplot" package; S44: According to the size of the correlation coefficient and the significance level (such as |R|>0.5 and p value<0.05), screen out LncRNAs significantly correlated with LRPPRC.

[0012] Preferably, in S5, the risk model construction and verification includes: S51: Based on the screened LncRNAs significantly correlated with LRPPRC, perform univariate Cox regression analysis using the "Survival" package to screen out LncRNAs with prognostic value; S52: Adopt LASSO regression analysis (using "glmnet" package) to optimize the model by least absolute shrinkage and selection operator method to reduce the risk of overfitting; S53: According to the LASSO regression results, construct a risk model composed of specific LncRNAs (such as CTD-2510F5.4 and SNHG4); S54: Calculate the risk score of each sample, and divide the patients into high-risk group and low-risk group according to the median risk score; S55: Draw time-dependent ROC curve using "timeROC" package to evaluate the sensitivity and specificity of the model; S56: Compare the survival rates of high-risk group and low-risk group by Kaplan-Meier survival curve to verify the prognostic prediction ability of the model.

[0013] Preferably, in S6, the independent prognostic analysis and nomogram construction include: S61: Collate and analyze the relationship between high and low risk groups of each sample and clinicopathological parameters (such as age, gender, grade, stage, etc.); S62: Perform univariate and multivariate Cox regression analysis using the "survival" R package to explore whether the risk score can be an independent biomarker for the survival of hepatocellular carcinoma patients; S63: Construct a nomogram integrating the risk score and various clinical characteristics to visually demonstrate the impact of different factors on prognosis; S64: Draw a calibration curve to evaluate the predictive accuracy of the nomogram and ensure that it accurately reflects the actual survival of patients.

[0014] Preferably, in S7, the functional enrichment analysis includes: S71: Perform Gene Ontology (GO) enrichment analysis on the differentially expressed genes in the high and low risk groups to identify significant changes in biological processes, cellular components, and molecular functions; S72: Perform Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis; S73: Analyze the correlation between cell division-related genes and LncRNAs (such as CTD-2510F5.4 and SNHG4) in the risk model.

[0015] Preferably, the m6A methylation regulatory factor is a leucine-rich PPR motif protein.

[0016] Preferably, the specific LncRNAs are CTD-2510F5.4 and SNHG4.

[0017] The beneficial effects of the present application are: 1. The present application combines the m6A methylation regulatory factor LRPPRC with LncRNAs to construct a hepatocellular carcinoma prognosis model, providing a new perspective and method for prognosis evaluation; through a combination of bioinformatics analysis and experimental verification, LRPPRC-related LncRNAs highly related to the survival and prognosis of hepatocellular carcinoma patients are screened; the constructed risk model shows excellent predictive performance in Kaplan-Meier survival curve and ROC curve analysis, which can accurately distinguish high and low risk groups and provide strong support for clinical decision-making.

[0018] 2. The present application uses univariate Cox regression analysis and LASSO regression analysis to further optimize the risk model; which can effectively reduce the risk of overfitting, improve the sensitivity and specificity of the model, and ensure the reliability of the model in practical application.

[0019] 3. The present application verifies the potential of the risk score as an independent biomarker for the survival of patients with hepatocellular carcinoma through independent prognostic analysis and nomogram construction, and the model is not affected by other confounding factors, and can more accurately reflect the prognosis of patients. BRIEF DESCRIPTION OF DRAWINGS

[0020] Figure 1 A schematic diagram for screening of m6A methylation regulatory factors and related long-chain non-coding RNAs of the present application; Figure 2 A heat map of the expression of 23 m6A methylation regulatory factors in 50 normal tissues and 374 tumor tissues of the present application; Figure 3 A multivariate Cox regression analysis forest plot of the prognostic value of 23 m6A methylation regulatory factors of the present application; Figure 4 A schematic diagram of the expression of m6A methylation regulatory factors in tumor tissues and normal tissues of the present application; Figure 5 An immunofluorescence analysis diagram of LRPPRC of the present application; Figure 6 A schematic diagram of hematoxylin-eosin staining and immunohistochemical staining of LRPPRC expression in paracancerous tissues and hepatocellular carcinoma tissues of the present application; Figure 7 A survival analysis diagram of LRPPRC of the present application; Figure 8 A correlation heat map of LRPPRC-related LncRNAs of the present application; Figure 9 A univariate Cox regression analysis prognostic value forest plot for identifying 5 related LncRNAs of the present application; Figure 10 A diagram of the minimum standard calculation of Lasso regression of the present application; Figure 11 A diagram showing the calculation coefficients of 2 related LncRNAs for constructing a risk model of the present application; Figure 12 A risk score distribution diagram of the present application; Figure 13 A survival status distribution diagram of the present application; Figure 14 A Kaplan-Meier survival curve diagram of the risk model of the present application; Figure 15 A receiver operating characteristic curve diagram showing the accuracy of the risk score in predicting 1-year, 2-year and 3-year survival rates of the present application; Figure 16 A univariate Cox regression analysis diagram showing that pathological stage, T stage, M stage and risk score are related to survival status. Figure 17 This is a schematic diagram illustrating how the multivariate Cox regression analysis of this invention shows that only the risk score can be used as an independent prognostic factor. Figure 18 This is a nomogram showing the predicted 1-year, 2-year, and 3-year survival rates according to the present invention. Figure 19 Volcano plot for differential expression analysis between the high-risk and low-risk groups in this invention; Figure 20 This is a diagram illustrating the enrichment status of GO enrichment analysis in this invention. Figure 21 This is a diagram showing the intersection analysis of the cell division genes of this invention with six enriched cell division pathways. Detailed Implementation

[0021] The technical solution of the present invention will be further described in detail below with reference to specific embodiments. Example

[0022] A prognostic model for hepatocellular carcinoma (HCC) was developed to explore the prognostic characteristics of LRPPRC-related long non-coding RNAs (lncRNAs) in order to predict the prognosis of HCC patients. The specific bioinformatics analysis route is shown in Figure 1. After obtaining LRPPRC-related lncRNAs with prognostic value, a risk model was established. Patients were divided into high-risk and low-risk groups based on median risk scores. The prognostic model was validated, and clinicopathological features were analyzed, and nomograms were constructed to verify its prognostic assessment capability. Finally, enrichment analysis of the risk model was performed using Gene Ontology (GO) and the Kyoto Encyclopedia of Genes and Genomes (KEGG), and further analysis of the enriched pathways was conducted; details are as follows: Materials and Methods: TCGA database We downloaded gene expression RNA sequencing data (HTSeq-FPKM and HTSeq-Count), clinical data, and survival data from 374 hepatocellular carcinoma (HCC) samples and 50 normal samples from the Cancer Genome Atlas (TCGA) database of UCSC Xena (https: / / xenabro.wser.net / datapages / ).

[0023] Identification and screening of m6A methylation regulators Twenty-three m6A methylation regulators were identified from published articles (Weng, C., et al., Identification of a N6-Methyladenosine (m6A)-Related lncRNA Signature for Predicting the Prognosis and Immune Landscape of Lung Squamous Cell Carcinoma. Frontiers in oncology, 2021. 11: p. 763027), and expression matrices, including the writers (METTL3, METTL14, METTL16, WTAP, KIAA1429, RBM15, RBM15B), were obtained from TCGA-LIHC expression data. Gene expression heatmaps were generated using the R language's "heatmap" package. Multivariate and univariate Cox regression analyses were performed using the R language's "survival" package to screen for the m6A methylation regulator LRPPRC (P < 0.01). Next, survival analysis was performed on the top 25% and bottom 25% of patients with LRPPRC expression using "GraphPadprism8", and differential expression between tumor tissue and adjacent normal tissue was visualized using the R language's "ggplot2" package.

[0024] Immunohistochemical analysis Tissue sample collection and sectioning were performed at the First Affiliated Hospital of Chongqing Medical University, China. This study was approved by the First Affiliated Hospital of Chongqing Medical University, China, and informed consent was obtained from all patients for the acquisition of tissue samples. Sections were baked at 75°C for 1 hour, followed by washing with xylene and graded ethanol. Endogenous peroxidase activity was blocked at room temperature with 3% hydrogen peroxide for 10 minutes. Antigen heat retrieval was performed in citrate buffer (pH=6.0). Then, 5% normal goat serum was added, and a blocking reaction was performed at 37°C for 30 minutes. Next, 50 μL of anti-LRPPRC antibody (1:20, Sangon Biotech, China) was added to these tissue sections, and incubation was carried out overnight at 4°C. After rewarming at 37°C for 30 minutes, the sections were incubated with secondary antibody at room temperature for 30 minutes. Finally, horseradish peroxidase was bound, and DAB was used for color development. The remaining steps were performed according to the kit instructions (SA1020, BOSTER, China).

[0025] Immunofluorescence analysis Cells were fixed with 4% neutral formaldehyde solution and then permeabilized with 0.3% Triton X-100. They were then blocked with 5% goat serum at 37°C for 30 minutes. Next, diluted anti-LRPPRC antibody (A3365, ABclonal, China) was added to the cells, and the cells were incubated overnight at 4°C. After warming, secondary antibody was added, and the cells were incubated at 37°C for 1 hour. Finally, DAPI working solution was added to stain the cell nuclei, and the cells were observed under a fluorescence microscope.

[0026] Correlation analysis between differentially expressed long non-coding RNAs (lncRNAs) and LRPPRC All 15,053 lncRNAs were extracted from the lncRNA annotation files of GENCODE (https: / / www.gencodegenes.org). Differential expression analysis and computation were performed using the "DESeq2" function in R (the screening threshold for differentially expressed lncRNAs was P < 0.05 and |log2Fold| > 1). Subsequently, Pearson correlation analysis was performed on the differentially expressed lncRNAs and LRPPRC using the "corrplot" package (the screening criteria were |R| > 0.5 and p < 0.05).

[0027] Construction and validation of LRPPRC-related LncRNA risk models First, univariate Cox regression analysis was performed using the "Survival" package to screen for LncRNAs associated with overall survival (OS) (P<0.01). LASSO regression analysis was then performed on the screened LncRNAs using the "glmnet" package, and a risk model was constructed. Risk scores were calculated based on the coefficients of the corresponding LncRNAs obtained from the LASSO regression. Risk score = coefficient 1 × LncRNA (1) + coefficient 2 × LncRNA (2) + coefficient 3 × LncRNA (3) + ... + coefficient i × LncRNA (i). Then, hepatocellular carcinoma (HCC) patients were divided into high-risk and low-risk groups based on the median risk score. Time-dependent ROC curves for 1, 2, and 3 years were plotted using the "timeROC" package, and the sensitivity and specificity of the risk model were assessed using the area under the curve (AUC). Subsequently, the relationships between the high- and low-risk groups and clinicopathological parameters (age, sex, grade, stage, T stage, M stage, N stage) for each sample were analyzed. We used the "survival" R package to perform univariate and multivariate Cox regression analyses to explore whether the risk score can serve as an independent biomarker for survival in patients with hepatocellular carcinoma.

[0028] Creation and Verification of Nodal Charts Using the packages "Hmisc", "grid", "lattice", "Formula", "ggplot2", "rms", "foreign", and "survival" in the R language, nomograms and calibration curves were generated based on risk scores and various clinical characteristics (age, sex, grade, stage, T stage, M stage, and N stage) to predict the 1-year, 2-year, and 3-year overall survival (OS) of hepatocellular carcinoma patients, and the predictive accuracy of the nomograms was evaluated.

[0029] Correlation analysis of enriched genes A gene list was obtained by retrieving "cell division" from the Gene Ontology database (http: / / geneontology.org / ). The intersection of this gene list with enriched genes for nuclear division, mitotic nuclear division, mitotic sister chromatid separation, sister chromatid separation, nuclear chromosome separation, and chromosome separation was calculated, and a Venn diagram was generated using the Bioladder website (https: / / www.bioladder.cn). Gene expression heatmaps were then generated using the R packages "heatmap", "limma", and "ggplot2". Pearson correlation analysis was performed on cell division genes and long non-coding RNAs (SNHG4, CTD-2510F5.4) using the R package "corrplot". Finally, the correlation between cell division genes and the model was analyzed using GraphPadPrism (version 8.3.0). Relevant genes were selected based on the screening criteria of p-value < 0.05 and |R| > 0.3, and box plots were generated using the R packages "limma" and "ggplot2".

[0030] Statistical methods All statistical analyses were performed using R software (version 4.1.2) and GraphPadPrism (version 9.5.1). A p-value < 0.05 was considered statistically significant (* indicates p < 0.05, ** indicates p < 0.01, *** indicates p < 0.001).

[0031] result: Identification of m6A methylation regulators with prognostic value This study included 374 patients and 50 healthy controls. Based on the literature, 23 m6A methylation regulators were selected, and their expression was preliminarily studied and presented in a heatmap. Figure 1 Further univariate Cox regression analysis identified regulatory factors with prognostic value: YTHDF1, YTHDF2, IGF2BP3, and LRPPRC (P<0.01). Figure 2 Multivariate Cox regression analysis identified four regulatory factors with p < 0.05: ZC3H13, YTHDF2, RBMX, and LRPPRC. Figure 3 Based on gene expression patterns in normal tissues and patients, LRPPRC is expressed at higher levels in hepatocellular carcinoma (HCC). Figure 4 Furthermore, LRPPRC expression was analyzed using immunofluorescence and immunohistochemistry. Immunofluorescence results confirmed the experimental expression of LRPPRC; red fluorescence represents LRPPRC, which was expressed at higher levels in tumor tissues. Figure 5Immunohistochemical staining showed that LRPPRC showed diffuse brownish-yellow or dark brown positive signals in the perinuclear or cytoplasmic regions of liver cancer cells, indicating that LRPPRC is highly expressed in tumor tissue. Figure 6 Kaplan-Meier survival analysis showed that the overall survival rate of the LRPPRC high expression group was lower than that of the low expression group. Figure 7 Taking all factors into consideration, this experiment selected LRPPRC as the main regulator of m6A methylation.

[0032] Construction and validation of LRPPRC-related long non-coding RNA (lncRNA) risk models First, differential expression analysis was performed on the obtained lncRNAs between normal and liver cancer tissues. Among 15053 lncRNAs, 2602 differentially expressed lncRNAs were identified. Subsequently, the correlation between LRPPRC and lncRNAs was calculated using the Pearson correlation coefficient, revealing five lncRNAs significantly associated with LRPPRC (…). Figure 8 Next, univariate Cox regression analysis was used to identify three LncRNAs with prognostic value from these five LncRNAs: RP11-466F5.8, CTD-2510F5.4, and SNHG4. Figure 9 LASSO regression analysis was performed based on these three identified lncRNAs. Figure 10 Two lncRNAs were ultimately identified and their correlation coefficients were calculated: CTD-2510F5.4 = 0.290214412548584, SNHG4 = 0.4792978874802. Figure 11 Based on these two lncRNAs and their corresponding coefficients, an LRPPRC-related lncRNA model was constructed: Risk score = (0.290214412548584 × CTD - 2510F5.4) + (0.4792978874802 × SNHG4). After the model was constructed, hepatocellular carcinoma (HCC) patients were divided into high-risk and low-risk groups according to the median risk score. Between the high-risk and low-risk groups, the expression levels of both lncRNAs were higher in the high-risk group. The risk score distribution and survival status of the two groups are shown in the figure. Figure 12 and 13 Kaplan-Meier survival curves showed a significant difference in survival rates between the high-risk and low-risk groups, with high-risk scores associated with poorer overall survival. Figure 14 The area under the ROC curve (AUC) values ​​for 1 year, 2 years, and 3 years were 0.7558, 0.6635, and 0.6461, respectively. Figure 15These results validate the accuracy and specificity sensitivity of the risk model.

[0033] Independent prognostic analysis and nomogram construction validation Chi-square test was used to analyze the differences in clinical characteristics between the high-risk and low-risk groups (Table 1). Univariate Cox regression analysis showed that stage, T stage, M stage, and risk score were all associated with overall survival (OS) in patients with hepatocellular carcinoma (HCC). Figure 16 Next, multivariate Cox regression analysis showed that only T stage and risk score were significantly associated with prognosis ( Figure 17 The results above indicate that the risk score has independent prognostic value, and the model can be used as an independent prognostic feature. Nonographs were constructed using risk status, age, gender, stage, T stage, M stage, and N stage to predict the incidence of overall survival at 1, 2, and 3 years. The results show that the risk level of this prognostic model has a certain predictive ability in the nonograph. Figure 18 ).

[0034] Enrichment analysis GO and KEGG enrichment analyses were performed on differentially expressed genes in the high-risk and low-risk groups to further explore potential functional pathways. First, differentially expressed gene analysis was conducted between the high-risk and low-risk groups. The volcano plot showed that, based on the criteria of P < 0.05 and |log2fold| > 1, 1971 differentially expressed genes were identified. Figure 19 Next, GO enrichment analysis was performed and visualized. Figure 20 GO enrichment analysis showed that biological processes (BP) were enriched in nuclear division, organelle division, and mitotic nuclear division; cellular components (CC) were enriched in the collagen-containing extracellular matrix and apical regions; and molecular functions (MF) were enriched in channel activity, passive transmembrane transporter activity, and gated channel activity. KEGG enrichment analysis showed that its potential functions include neuroactive ligand-receptor interactions, retinol metabolism, and drug metabolism-cytochrome P450.

[0035] Correlation analysis between risk models and enriched genes Because a large number of cell division pathways were enriched, the relationship between cell division pathways and the model was further analyzed. Cell division genes were obtained by searching for "cell division" in the gene ontology database (http: / / geneontology.org / ). First, the obtained cell division genes were subjected to intersection analysis with the genes of six enriched cell division pathways: "nuclear division," "sister chromatid separation," "chromosome separation," "mitotic nuclear division," "mitotic sister chromatid separation," and "nuclear chromosome separation," resulting in 46 common genes. Figure 21Analysis of the expression of these 46 genes revealed that most were highly expressed in the high-risk group with strong statistical significance. In addition, 43 genes were associated with CTD-2510F5.4 and SNHG4 in the model, and these genes were also highly expressed in tumor tissues. Correlation analysis showed that these 43 genes were significantly correlated with the risk score, with the nine genes exhibiting the highest correlation. These results indicate a close association between this risk model and cell division.

[0036] The relationship between prognostic models and clinicopathological features of HCC is as follows: Case High Risk of the model(%) 150(50.1%) Low Risk of the model(%) 149(49.8%) P-value Age 0.203 ≥60 year 157 68（22.7） 60（20.1） <60 year 142 82（27.4） 89（29.8） Gender 0.028 Female 95 57（19.1） 38（12.7） Male 204 93（31.1） 111（37.1） Grade <0.001 G1 41 13（4.3） 28（9.4） G2 139 59（19.7） 80（26.8) G3 107 70（23.4） 37（12.4） G4 12 8（2.7） 4（1.3） Stage 0.037 Stage1 153 66（22.1） 87（29.1） Stage2 71 39（13） 32（10.7) Stage3 75 45（15.1） 30（10.0） T 0.19 T1 151 67（22.4） 84（28.1） T2 72 40（13.4） 32（10.7） T3 67 39（13） 28（9.4） T4 9 4（1.3） 5（1.7） M 0.139 M1 3 0（0.0） 3（1.0） M0 232 121（40.5） 111（37.1） MX 64 29（9.7） 35（11.7） N 0.101 N0 227 117（39.1） 110（36.8） N1 3 3（1.0） 0（0.0） NX 69 30（10.0） 39（13.0） Where T represents the primary pathological tumor; m represents pathological metastasis; and n represents the pathological lymph node status.

[0037] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A method for constructing a prognostic model for hepatocellular carcinoma, characterized in that, Includes the following steps: S1: Data acquisition and preprocessing; S2: Identification and screening of m6A methylation regulators; S3: Immunohistochemistry and immunofluorescence analysis; S4: Correlation analysis between differentially expressed long non-coding RNAs and LRPPRC; S5: Risk model construction and validation; S6: Independent prognostic analysis and nomogram construction; S7: Functional enrichment analysis.

2. The method for constructing a hepatocellular carcinoma prognostic model according to claim 1, characterized in that, In step S1, data acquisition and preprocessing include: S12: Download RNA sequencing data, clinical information, and survival data of hepatocellular carcinoma patients and normal controls from the UCSC Xena database; S13: Clean the data and remove invalid or missing data entries; S14: Standardize the data to ensure consistency across different sources; S15: Handle missing values ​​and fill in the missing data using imputation methods.

3. The method for constructing a hepatocellular carcinoma prognostic model according to claim 1, characterized in that, In S2, the identification and screening of m6A methylation regulators includes: S21: Determine the list of m6A methylation regulators; S22: Extract the expression matrix of these regulatory factors in hepatocellular carcinoma patients and normal controls; S23: Use the heatmap package in R language to draw gene expression heatmaps to show the expression of regulatory factors in different samples; S24: Using the survival package in R language, univariate Cox regression analysis was performed to preliminarily screen out regulatory factors related to survival; S25: Perform multivariate Cox regression analysis to screen regulatory factors with independent prognostic value.

4. The method for constructing a hepatocellular carcinoma prognostic model according to claim 1, characterized in that, In S3, immunohistochemistry and immunofluorescence analysis include: S31: Collect tissue samples from patients with hepatocellular carcinoma; S32: Section the tissue sample and perform fixation and dehydration treatment; S33: Immunohistochemical staining was performed, and tumor cells were labeled with anti-LRPPRC antibody. The expression pattern of LRPPRC was observed and recorded under a microscope. S34: Perform immunofluorescence analysis using anti-LRPPRC antibody, combined with fluorescent markers, and observe and record the expression of LRPPRC under a fluorescence microscope.

5. The method for constructing a hepatocellular carcinoma prognostic model according to claim 1, characterized in that, In S4, the correlation analysis between differentially expressed long non-coding RNAs and LRPPRC includes: S41: Obtain annotation files for all LncRNAs from the GENCODE database; S42: Differential expression analysis was performed using the DESeq2 package in R language to identify LncRNAs that were significantly differentially expressed in hepatocellular carcinoma; S43: Calculate the Pearson correlation coefficient between differentially expressed LncRNAs and LRPPRC using the corrplot package; S44: Based on the magnitude and significance level of the correlation coefficient, screen out LncRNAs that are significantly associated with LRPPRC.

6. The method for constructing a hepatocellular carcinoma prognostic model according to claim 1, characterized in that, In S5, the risk model construction and validation includes: S51: Based on the LncRNAs that have been screened and are significantly associated with LRPPRC, univariate Cox regression analysis was performed using the Survival package to screen out LncRNAs with prognostic value. S52: LASSO regression analysis was used, and the model was optimized by the minimum absolute shrinkage and selection operator method to reduce the risk of overfitting; S53: Construct a risk model composed of specific LncRNAs based on LASSO regression results; S54: Calculate the risk score for each sample and classify patients into high-risk and low-risk groups based on the median risk score; S55: Use the timeROC package to plot the time-dependent ROC curve to evaluate the model's sensitivity and specificity; S56: Compare the survival rates of the high-risk group and the low-risk group using Kaplan-Meier survival curves to verify the model's prognostic predictive ability.

7. The method for constructing a hepatocellular carcinoma prognostic model according to claim 1, characterized in that, In step S6, independent prognostic analysis and nomogram construction include: S61: Compile and analyze the relationship between the high-risk and low-risk groups and clinicopathological parameters for each sample; S62: Use the survival R package for univariate and multivariate Cox regression analysis; S63: Construct nomograms to integrate risk scores and various clinical characteristics, and visualize the impact of different factors on prognosis. S64: Plot calibration curves to assess the predictive accuracy of nomograms and ensure they accurately reflect the patient's actual survival.

8. The method for constructing a hepatocellular carcinoma prognostic model according to claim 1, characterized in that, In S7, the functional enrichment analysis includes: S71: Perform gene ontology enrichment analysis on differentially expressed genes in high-risk and low-risk groups to identify significant changes in biological processes, cellular components and molecular functions. S72: Perform enrichment analysis of the Kyoto Encyclopedia of Genes and Genomes; S73: Analyze the correlation between cell division-related genes and LncRNAs in the risk model.

9. The method for constructing a hepatocellular carcinoma prognostic model according to claim 3, characterized in that, The m6A methylation regulator is a leucine-rich PPR motif protein.

10. The method for constructing a hepatocellular carcinoma prognostic model according to claim 6, characterized in that, The specific LncRNAs mentioned are CTD-2510F5.4 and SNHG4.

Images