Background. Hepatocellular carcinoma (HCC) remains an important cause of cancer death. The molecular mechanism of hepatocarcinogenesis and prognostic factors of HCC have not been completely uncovered. Methods. In this study, we screened out differentially expressed lncRNAs (DE lncRNAs), miRNAs (DE miRNAs), and mRNAs (DE mRNAs) by comparing the gene expression of HCC and normal tissue in The Cancer Genome Atlas (TCGA) database. DE mRNAs were used to perform Gene Ontology (GO), Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis. Then, the miRNA and lncRNA/mRNA modules that were most closely related to the survival time of patients with HCC were screened to construct a competitive endogenous RNA (ceRNA) network by weighted gene coexpression network analysis (WGCNA). Moreover, univariable Cox regression and Kaplan-Meier curve analyses of DE lncRNAs and DE mRNAs were conducted. Finally, the lasso-penalized Cox regression analysis and nomogram model were used to establish a new risk scoring system and predict the prognosis of patients with liver cancer. The expression of survival-related DE lncRNAs was verified by qRT-PCR. Results. A total of 1896 DEmRNAs, 330 DElncRNAs, and 76 DEmiRNAs were identified in HCC and normal tissue samples. Then, the turquoise miRNA and turquoise lncRNA/mRNA modules that were most closely related to the survival time of patients with HCC were screened to construct a ceRNA network by WGCNA. In this ceRNA network, there were 566 lncRNA-miRNA-mRNA regulatory pairs, including 30 upregulated lncRNAs, 16 downregulated miRNAs, and 75 upregulated mRNAs. Moreover, we screened out 19 lncRNAs and 14 hub mRNAs related to prognosis from this ceRNA network by univariable Cox regression and Kaplan-Meier curve analyses. Finally, a new risk scoring system was established by selecting the optimal risk lncRNAs from the 19 prognosis-related lncRNAs through lasso-penalized Cox regression analysis. In addition, we established a nomogram model consisting of independent prognostic factors to predict the survival rate of HCC patients. Finally, the correlation between the risk score and immune cell infiltration and gene set enrichment analysis were determined. Conclusions. In conclusion, the results may provide potential biomarkers or therapeutic targets for HCC and the establishment of the new risk scoring system and nomogram model provides the new perspective for predicting the prognosis of HCC.

中文翻译：

---

基于WGCNA筛选预后相关lncRNA，建立预测肝细胞癌患者预后的新风险评分


背景。肝细胞癌（HCC）仍然是癌症死亡的重要原因。肝癌发生的分子机制和HCC的预后因素尚未完全阐明。方法。在本研究中，我们通过比较癌症基因组图谱（TCGA）数据库中HCC和正常组织的基因表达量，筛选出差异表达的lncRNAs（DE lncRNAs）、miRNAs（DE miRNAs）和mRNAs（DE mRNAs）。 DE mRNA 用于进行基因本体论 (GO)、京都基因和基因组百科全书 (KEGG) 通路分析。然后，通过加权基因共表达网络分析（WGCNA）筛选与HCC患者生存时间最密切相关的miRNA和lncRNA/mRNA模块，构建竞争性内源RNA（ceRNA）网络。此外，还对 DE lncRNA 和 DE mRNA 进行了单变量 Cox 回归和 Kaplan-Meier 曲线分析。最后，利用套索惩罚Cox回归分析和列线图模型建立新的风险评分系统并预测肝癌患者的预后。通过 qRT-PCR 验证了与生存相关的 DE lncRNA 的表达。结果。在 HCC 和正常组织样本中总共鉴定出 1896 个 DEmRNA、330 个 DElncRNA 和 76 个 DEmiRNA。然后，通过WGCNA筛选与HCC患者生存时间最密切相关的绿松石miRNA和绿松石lncRNA/mRNA模块，构建ceRNA网络。在这个ceRNA网络中，有566个lncRNA-miRNA-mRNA调控对，包括30个上调的lncRNA、16个下调的miRNA和75个上调的mRNA。 此外，我们通过单变量Cox回归和Kaplan-Meier曲线分析，从该ceRNA网络中筛选出与预后相关的19个lncRNA和14个hub mRNA。最后，通过套索惩罚Cox回归分析从19个预后相关lncRNA中选择最佳风险lncRNA，建立了新的风险评分系统。此外，我们建立了由独立预后因素组成的列线图模型来预测 HCC 患者的生存率。最后，确定风险评分与免疫细胞浸润之间的相关性以及基因集富集分析。结论。总之，该结果可能为HCC提供潜在的生物标志物或治疗靶点，新的风险评分系统和列线图模型的建立为预测HCC的预后提供了新的视角。