Qu Yaqi, Chen Yujia, Zhang Le, Tian Lifei
Department of General Surgery, Shaanxi Provincial People's Hospital, Xi'an, 710068, China.
School of Basic Medical Sciences, Hebei Medical University, Shijiazhuang, Hebei, China.
World J Surg Oncol. 2020 Sep 3;18(1):236. doi: 10.1186/s12957-020-02010-7.
Colon adenocarcinoma (COAD) is one of the most common malignant tumors, with high incidence and mortality rates worldwide. Reliable prognostic biomarkers are needed to guide clinical practice.
Comprehensive gene expression with alternative splicing (AS) profiles for each patient was downloaded using the SpliceSeq database from The Cancer Genome Atlas. Cox regression analysis was conducted to screen for prognostic AS events. The R package limma was used to screen differentially expressed genes (DEGs) between normal and tumor samples in the COAD cohort. A Venn plot analysis was performed between DEGs and prognostic AS events, and the DEGs that co-occurred with prognostic AS events (DEGAS) were identified. The top 30 most-connected DEGAS in protein-protein interaction analysis were identified through Cox proportional hazards regression to establish prognostic models.
In total, 350 patients were included in the study. A total of 22,451 AS events were detected, of which 2004 from 1439 genes were significantly associated with survival time. By overlapping these 1439 genes with 6455 DEGs, 211 DEGs with AS events were identified. After the construction of the protein-protein interaction network, the top 30 hub genes were included in a multivariate analysis. Finally, a risk score based on 12 genes associated with overall survival was established (P < 0.05). The area under the curve was 0.782. The risk score was an independent predictor (P < 0.001).
By exploring survival-associated AS events, a powerful prognostic predictor consisting of 12 DEGAS was built. This study aims to propose a novel method to provide treatment targets for COAD and guide clinical practice in the future.
结肠腺癌（COAD）是最常见的恶性肿瘤之一，在全球范围内发病率和死亡率都很高。需要可靠的预后生物标志物来指导临床实践。
使用来自癌症基因组图谱的SpliceSeq数据库下载每位患者的具有可变剪接（AS）图谱的综合基因表达数据。进行Cox回归分析以筛选预后性AS事件。使用R包limma筛选COAD队列中正常样本和肿瘤样本之间的差异表达基因（DEG）。对DEG和预后性AS事件进行韦恩图分析，确定与预后性AS事件同时出现的DEG（DEGAS）。通过Cox比例风险回归确定蛋白质-蛋白质相互作用分析中连接性最强的前30个DEGAS，以建立预后模型。
该研究共纳入350例患者。共检测到22451个AS事件，其中来自1439个基因的2004个与生存时间显著相关。通过将这1439个基因与6455个DEG重叠，确定了211个具有AS事件的DEG。构建蛋白质-蛋白质相互作用网络后，将前30个枢纽基因纳入多变量分析。最后，建立了一个基于12个与总生存相关基因的风险评分（P < 0.05）。曲线下面积为0.782。风险评分是一个独立的预测指标（P < 0.001）。
通过探索与生存相关的AS事件，构建了一个由12个DEGAS组成的强大预后预测指标。本研究旨在提出一种新方法，为COAD提供治疗靶点并指导未来的临床实践。
