Gastroenterology Department, Beijing Aerospace General Hospital, Beijing, China.
J Appl Genet. 2023 Dec;64(4):713-721. doi: 10.1007/s13353-023-00778-4. Epub 2023 Aug 17.
The objective of this study was to identify methylation-driven genes and explore their prognostic value in colon adenocarcinoma (COAD). The Cancer Genome Atlas (TCGA) database was used to acquire collated COAD transcriptome gene expression matrix (containing 59,427 transcripts), transcriptome gene methylation level matrix (containing 29,602 methylated modified genes), which included 517 samples containing 41 samples of normal tissue (NT) & 476 samples of COAD, and patient clinical information files (including patient survival time, survival status, age, gender and tumor stage, etc.), for all COAD samples. A total of 9807 differentially expressed genes (DEGs) were obtained by DEG analysis of the COAD transcriptional expression matrix, of which 5874 were up-regulated and 3933 were down-regulated. And 46 methylation-driven DEGs (MD-DEGs) in COAD were obtained by DEG analysis, differential analysis of gene methylation levels, and correlation analysis between them. Next, three prognostic associated MD-DEGs (PMD-DEGs) (IDUA, ZBTB18 and C5orf38) were identified by Cox regression analysis, and a prognostic model composed of the three PMD-DEGs was constructed by least absolute shrinkage and selection operator (LASSO) regression analysis and cross-validation analysis. In addition, survival analysis, the receiver operating characteristics (ROC) curve analysis and independent prognostic analysis were used to evaluate and verify that the prognostic model we constructed could accurately and independently predict the prognosis of COAD patients. Finally, we constructed a nomogram based on the prognosis model to accurately and personalized predict the survival prognosis of COAD patients. In conclusion, we identified the methylation driver gene of COAD and constructed a prognostic model and nomogram to personalized predict the prognosis of patients, which opened a new prospect for accurate diagnosis and treatment in clinical practice.
本研究旨在鉴定甲基化驱动基因，并探讨其在结肠腺癌（COAD）中的预后价值。我们使用癌症基因组图谱（TCGA）数据库获取了 COAD 转录组基因表达矩阵（包含 59427 个转录本）和转录组基因甲基化水平矩阵（包含 29602 个甲基化修饰基因），其中包含 517 个样本，包括 41 个正常组织（NT）样本和 476 个 COAD 样本，以及患者临床信息文件（包括患者生存时间、生存状态、年龄、性别和肿瘤分期等）。通过对 COAD 转录表达矩阵进行 DEG 分析，共获得 9807 个差异表达基因（DEGs），其中上调基因 5874 个，下调基因 3933 个。通过对 COAD 基因甲基化水平的 DEG 分析、差异分析和相关性分析，共获得 46 个甲基化驱动的 DEGs（MD-DEGs）。然后，通过 Cox 回归分析筛选出 3 个预后相关的 MD-DEGs（PMD-DEGs）（IDUA、ZBTB18 和 C5orf38），并通过最小绝对收缩和选择算子（LASSO）回归分析和交叉验证分析构建了由这 3 个 PMD-DEGs 组成的预后模型。此外，我们还进行了生存分析、受试者工作特征（ROC）曲线分析和独立预后分析，以评估和验证我们构建的预后模型能够准确、独立地预测 COAD 患者的预后。最后，我们基于预后模型构建了列线图，以准确、个性化地预测 COAD 患者的生存预后。总之，我们鉴定了 COAD 的甲基化驱动基因，并构建了预后模型和列线图，以个性化预测患者的预后，为临床实践中的精准诊断和治疗开辟了新的前景。
