基于DNA甲基化模式筛选评估结肠腺癌预后的标记物

为了通过分析DNA甲基化谱识别出与预后相关的结肠腺癌亚型。从TCGA数据库获取了结肠腺癌患者的甲基化数据,通过差异甲基化分析和构建COX比例风险回归模型筛得与预后显著相关的CpG位点,并通过一致性聚类识别出7个亚型。生存分析和临床特征检验显示7个亚型间预后差异显著且亚型特征可由多种临床特征反映。此外,用7个亚型间识别出的差异甲基化位点构建的基于SMO(序列最小最优化)的预测模型在各亚型上都有较高的AUC值,并用检验集进行了验证。综上,本研究利用生物信息学算法识别了7个预后差异的结肠腺癌亚型并挖掘了它们的特异性甲基化标记。该研究结果或可使得结肠腺癌预后被更精准地评估,为早期诊断及治疗方案提供新思路。

Marker evaluation for the prognosis of colon adenocarcinoma based on DNA methylation patterns

To identify the subtypes of colon adenocarcinoma for prognosis by analyzing the DNA methylation patterns,methylation data of colon adenocarcinoma patients was obtained from the TCGA database, and CpG sites that significantly related to prognosis were screened by differential methylation analysis and constructing a COX proportional hazard regression model. Seven subtypes were identified through consensus clustering. Survival analysis and clinical feature tests showed that the prognosis was significantly different among the seven subtypes and the subtype features could be reflected by multiple clinical features. In addition, the sequence minimal optimization(SMO)-based prediction model was constructed based on the differential methylation data and showed high AUC values for each subtype. Finally, the SMO model was evaluated on the test set. To sum up, this study used bioinformatics algorithms to identify seven subtypes of colon adenocarcinoma with different prognosis and their specific methylation markers. The results of this study enable the prognosis of colon adenocarcinoma to be more accurately evaluated, providing new ideas for early diagnosis and treatment.