可变剪接对肾透明细胞癌患者预后的预测价值

Predictive value on prognosis of alternative splicing in patients with renal clear cell carcinoma

目的利用生物信息学方法研究可变剪接对肾透明细胞癌(RCCC)患者预后的预测价值。方法从癌症基因组图谱(TCGA)数据库下载RCCC患者可变剪接事件相关RNA-seq数据及其临床信息,通过单因素Cox回归分析筛选与生存显著相关的可变剪接事件。将选择的可变剪接事件进行LASSO回归分析,并构建Cox预测模型。结果 497例RCCC患者癌组织的21 615个基因中共有46 415个可变剪接事件发生,单因素Cox回归分析结果显示,在5 483个基因中发现12 156个与RCCC患者生存显著相关的可变剪接事件(均P <0.05)。LASSO回归分析结果显示,可变受体位点(AA)事件中与RCCC患者生存预后相关性最高的基因为6个,可变供体位点(AD)、可变启动子(AP)、可变终止子(AT)、外显子跳跃(ES)事件中均为5个,外显子互斥(ME)事件中为10个,内含子保留(RI)事件中为4个,整体可变剪接事件中为6个。将7种可变剪接事件及整体可变剪接事件分别构建Cox预测模型,Kaplan-Meier生存曲线分析显示各预测模型中高风险组生存率均低于低风险组(均P <0.001)。受试者工作特征曲线结果显示,AA、AD、AP、AT、ES、ME、RI和整体可变剪接预测模型的曲线下面积分别为0.759、0.753、0.743、0.763、0.789、0.762、0.674和0.796。结论可变剪接事件构建的预测模型能有效地预测RCCC患者的生存情况,为临床治疗决策提供帮助。

Objective To investigate predictive value on prognosis of alternative splicing( AS) in patients with renal clear cell carcinoma( RCCC) by bioinformatics methods. Methods The RNA-seq data and clinical information related to AS events in patients with RCCC were downloaded from the Cancer Genome Atlas( TCGA)database,and the AS events that were significantly related to survival were screened by univariate Cox regression analysis. The selected AS events were analyzed by LASSO regression analysis and were used to construct Cox predictive model. Results Totally 46 415 AS events were expressed among 21 615 genes in cancer tissue of 497 patients with RCCC. Univariate Cox regression analysis showed that 12 156 AS events were found to be significantly related to survival among 5 483 genes( all P < 0. 05). LASSO regression analysis showed that there were 6 genes mostly related to survival prognosis of patients with RCCC in alternative acceptor( AA) event,5 genes in alternative donor( AD),alternative promoter( AP),alternative terminator( AT) and exon skipping( ES) events,10 genes in mutual exclusive exon( ME) event,4 genes in retained intron( RI) event,and 6 genes in all-AS events. Seven kinds of AS events and all-AS events were used to construct Cox predictive model,respectively. Kaplan-Meier survival curve analysis showed that the survival rate of high-risk group was lower than that of low-risk group in each predictive model( all P < 0. 001). The receiver operating characteristic curve showed that,the areas under the curve of AA,AD,AP,AT,ES,ME,RI and all-AS predictive models were 0. 759,0. 753,0. 743,0. 763,0. 789,0. 762,0. 674 and 0. 796,respectively. Conclusion The predictive model constructed with AS events can effectively predict the survival of patients with RCCC and provide help for clinical treatment decisions.