结肠癌免疫相关 lncRNA 风险预测模型的建立

目的 分析影响结肠癌预后的免疫相关 lncRNA, 并构建预测结肠癌患者预后的相关预测模型。 方法 下载 TCGA 数据库中的结肠癌 lncRNA 表达谱, 数据经 TPM 标准化后分析所有 lncRNA 的差异性表 达, 其中缺失值的补充采用 KNN 法, 通过共表达方法提取并鉴定免疫相关 lncRNA, 然后对差异性表达的 lncRNA 进行 LASSO 回归分析, 再进行单因素和多因素 COX 回归分析。 最后使用 R 4. 0. 2 统计学软件的 ggplot2 包基于 lncRNA 风险评分与基因表达关系, 构建风险因子关联图、 KM 曲线及评价模型预测价值的 ROC 曲线。 结果 经过表达差异性分析发现, 共有 2 258 个 lncRNA 在癌和癌旁组织中差异性表达, 其中上 调的有 1 648 个, 下调的有 610 个。 选取差异表达前 100 位的免疫相关 lncRNA 进行 LASSO 回归分析, 共筛 选出 12 个 lncRNA, 再进行单因素和多因素 COX 回归分析后显示 AC092723. 1、 AC007182. 1 和 AC004947. 1 与预后明显相关, 使用 R 4. 0. 2 统计学软件构建预后风险因子关联图, ROC 曲线显示其预测 1 年、 3 年和 5 年的预测价值均较高, 其 AUC 分别为 0. 79 (95 % CI: 0. 67 ~ 0. 91), 0. 78 (95 % CI: 0. 66 ~ 0. 9), 0. 7 (95 % CI: 0. 51 ~ 0. 9)。 结论 研究使用 TCGA 公共数据库进行生物信息学分析并构建的预后模型显示有 较高的预测价值, 除具有一定的临床意义外, 对未来 lncRNA 相关结肠癌的研究也提供了一定的方向。

Establishment of Immune Related lncRNA Risk Prediction Model for Colon Cancer

Objective To analyze the immune related lncRNAs that affect the prognosis of colon cancer, and construct a related prediction model for the prediction of the prognosis of colon cancer patients. Methods Download the lncRNA expression profile of colon cancer in the TCGA database, and the data was standardized by TPM to analyze the differential expression of all lncRNAs. The KNN method was used to supplement the missing values. Immune-related lncRNAs were extracted and identified of by co-expression method, and then LASSO regression analysis was performed on the top 100 differentially expressed lncRNAs. Then the single-factor and multi-factor COX regression analysis were performed. Finally, based on the relationship between the lncRNA risk score and the gene expression, the risk factor association chart, the KM curve and the ROC curve for the evaluation of the value of the predicted model were constructed by using the ggplot2 package of R 4. 0. 2 statistical software. Results Through the analysis of differencial expression, a total of 2258 lncRNAs were found to be differentially expressed in the cancer and paracancerous tissues, of which 1 648 were up-regulated and 610 were down-regulated. The top 100 differentially expressed immune related lncRNAs were selected for LASSO regression analysis, and a total of 12 lncRNAs were screened out. Univariate and multivariate COX regression analysis showed that AC092723. 1, AC007182. 1 and AC004947. 1 were significantly related to the prognosis. Using R 4. 0. 2 statistical software to construct a prognostic risk factor association chart. The ROC curve showed that the predictive values for the prediction of 1, 3 and 5 years were high, and the AUC were 0. 79 (95 % CI: 0. 67-0. 91), 0. 78 (95 % CI: 0. 66-0. 9), 0. 7 (95 % CI: 0. 51-0. 9), respectively. Conclusion This study uses the TCGA public database for the bioinformatics analysis and constructs a prognostic model that shows high predictive values. In addition to certain clinical significance, it also provides certain directions for the future researches of lncRNAs in colon cancer.