构建胃癌lncRNA预测模型以促进胃癌的治疗

Constructing lncRNA Prediction Model of Gastric Cancer to Promote the Treatment of Gastric Cancer

摘要: 目的：基于大样本基因测序数据构建胃癌的lncRNA预测模型。方法：下载GEO数据库和TCGA数据库中胃癌的基因测序数据和临床数据。应用单因素Cox回归筛选胃癌预后相关的lncRNA。应用机器学习-LASSO回归构建lncRNA预后模型。基于临床特征和lncRNA预测模型进行单因素Cox回归和多因素Cox回归以确定该模型为胃癌的独立预后因素。应用cibersort分析胃癌中免疫细胞的情况，并分析lncRNA预测模型及其构成基因与免疫细胞浸润的关系。结果：我们总共28个预后相关lncRNA。基于预后相关lncRNA，构建了一个由20个lncRNA组成的预测模型。应用外部验证集验证该模型的结果表明该lncRNA预测模型具有广泛适用性。通过分析lncRNA预测模型与临床特征的关系发现lncRNA预测模型与肿瘤大小和淋巴结转移密切相关。另外，我们发现该lncRNA预测模型可区分胃癌患者免疫细胞浸润情况。总结：总之，我们构建了一个lncRNA预测模型，该模型能够区分胃癌的淋巴结转移、肿瘤大小和免疫细胞浸润的状态。

Abstract: Objective: To construct the lncRNA prediction model for gastric cancer based on large sample gene sequencing data. Methods: The gene sequencing data and clinical data of gastric cancer were down-loaded from GEO database and TCGA database. Single factor Cox regression was used to screen lncRNA associated with the prognosis of gastric cancer. lncRNA prognostic model was constructed by machine learning-LASSO regression. Univariate Cox regression and multivariate Cox regression were performed based on clinical features and the lncRNA prediction model to determine that the model was an independent prognostic factor for gastric cancer. CIBERSORT was applied to analyze the situation of immune cells in gastric cancer, and the relationship between the lncRNA prediction model and its constituent genes and immune cell infiltration was analyzed. Results: We identified a total of 28 prognostic relevant lncRNAs. Based on prognostic related lncRNA, a prediction model consisting of 20 lncRNAs was constructed. The results of the external validation set show that the lncRNA prediction model has wide applicability. By analyzing the relationship between the lncRNA prediction model and clinical features, it was found that the lncRNA prediction model was closely related to tumor size and lymph node metastasis. In addition, we found that this lncRNA prediction model could distinguish immune cell infiltration in gastric cancer patients. Conclusions: In sum-mary, we constructed an lncRNA predictive model that was able to distinguish between lymph node metastasis, tumor size, and the state of immune cell invasion in gastric cancer.