结直肠癌失巢凋亡相关基因的生物信息学筛选及构建预后模型

Bioinformatics screening and construction of prognostic model of genes associated with anoikis in colorectal cancer

目的基于生物信息学探索失巢凋亡相关基因(anoikis-related genes,ANRGs)在结直肠癌(colorectal cancer,CRC)中的预后价值,并构建了基于ANRGs的预后模型。方法利用GeneCard和Harmonizome整合了515个ANRGs。首先从TCGA数据库获得CRC的mRNA表达数据,并利用“Limma”包获得ANRGs,采用单因素Cox回归筛选出与总生存期有关的ANRGs,再采用LASSO回归和多因素Cox回归构建ANRGs的预后模型。最后用ROC曲线分析来评估模型的预测性能,并使用GSE76427数据集进行验证。进一步分析风险评分与肿瘤免疫微环境相关性。将风险评分与临床病理特征相结合,建立基于ANRGs的列线图来预测总生存期。决策曲线分析法(decision curve analysis,DCA)评估列线图的效能。结果我们鉴定了102个与生存相关的CRC的ANRGs,并且选择了6个基因来构建预后风险评分模型。训练集(AUC max=0.725)和验证集(AUC max=0.629)证明该模型对患者的预后显示出较好的预测性能,预后风险评分被确定为独立的预后因素。功能分析表明,高危组和低危组具有不同的免疫浸润状态。DCA分析表明列线图可以从临床治疗策略中获益。结论本研究建立了一个由6个ANRGs组成的CRC预后模型,可以帮助临床医师为CRC患者选择个性化治疗提供重要参考价值。

Objective To explore the prognostic value of anoikis-related genes(ANRGs)in colorectal cancer(CRC)based on bioinformatics and construct a prognostic model based on ANRGs.Methods A total of 515 ANRGs were integrated from GeneCard and Harmonizome.Firstly,mRNA expression data of CRC was obtained from the TCGA database,and ANRGs were obtained using the"Limma"package.Univariate Cox regression was used to select ANRGs related to overall survival,and LASSO regression and multivariate Cox regression were used to construct a prognostic model for ANRGs.ROC curve analysis was performed to evaluate the predictive performance of the model,and validation was conducted using the GSE76427 dataset.Further analysis was conducted to explore the correlation between risk scores and tumor immune microenvironment.A column chart based on ANRGs was constructed to predict overall survival by combining risk scores with clinical pathological features.The decision curve analysis(DCA)was used to evaluate the efficacy of the column chart.Results We identified 102 ANRGs of survivorship associated CRC and selected 6 genes to construct a prognostic risk score model.The training set(AUC max=0.725)and validation set(AUC max=0.629)proved that the model showed good predictive performance for patients'prognosis,and the prognostic risk score was identified as an independent prognostic factor.Functional analysis showed that high risk group and low risk group had different immune infiltrating states.DCA analysis suggested that the Nomogram can benefit from clinical treatment strategies.Conclusion In this study,a CRC prognostic model consisting of 6 ANRGs was established,which can help clinicians to select personalized treatment for CRC patients with important reference value.