Development and validation of a 14-gene signature for prognosis prediction in hepatocellular carcinoma
Development and validation of a 14-gene signature for prognosis prediction in hepatocellular carcinoma
复制标题
用于肝细胞癌预后预测的 14 基因特征的开发和验证
DOI：
10.1016/j.ygeno.2020.03.013
复制
发表时间：
2020
期刊：
影响因子：
4.4
通讯作者：
Yang Jia-Yin
中科院分区：
文献类型：
--
作者：
Zhang Bo-Han;Yang Jian;Jiang Li;Lyu Tao;Kong Ling-Xiang;Tan Yi-Fei;Li Bo;Zhu Yun-Feng;Xi Ao-Yao;Xu Xi;Yan Lyu-Nan;Yang Jia-Yin
在世界范围内，肝细胞癌（HCC）仍然是一个重要的医学问题。由于肝癌细胞间复杂的基因变异，迫切需要精确而简洁的预后模型。我们进行了这项研究，以确定具有生物学意义的预后基因签名。我们应用了两种算法来生成癌症基因组图谱队列（包括训练集）中HCC和正常标本之间的差异表达基因（DEG），并进行富集分析以阐述其生物学意义。基于STRING在线工具建立了蛋白质相互作用网络。然后，我们使用Cytoscape筛选关键模块中的枢纽基因。通过枢纽基因的考克斯回归分析构建多基因签名，以在训练集中对HCC患者的疾病进行分层。多基因标签的预后价值在来自基因表达综合数据库（Gene Expression Omnibus）的另外两组（GSE 14520和GSE 76427）中进行了外部验证，并且还研究了其在复发预测中的作用。共获得2000个DEG，其中上调基因1542个，下调基因458个。随后，我们在56个枢纽基因的基础上构建了一个14基因签名，这是一个很好的总生存率预测因子。预后标志可以在GSE 14520和GSE 76427中复制。此外，14-基因标签可以应用于训练集和GSE 14520中的复发预测。总之，从枢纽基因中提取的14个基因签名参与了一些HCC相关的信号通路;它不仅可以作为HCC结果的预测签名，而且还可以用于预测HCC复发。
Worldwide, hepatocellular carcinoma (HCC) remains a crucial medical problem. Precise and concise prognostic models are urgently needed because of the intricate gene variations among liver cancer cells. We conducted this study to identify a prognostic gene signature with biological significance. We applied two algorithms to generate differentially expressed genes (DEGs) between HCC and normal specimens in The Cancer Genome Atlas cohort (training set included) and performed enrichment analyses to expound on their biological significance. A protein-protein interactions network was established based on the STRING online tool. We then used Cytoscape to screen hub genes in crucial modules. A multigene signature was constructed by Cox regression analysis of hub genes to stratify the prognoses of HCC patients in the training set. The prognostic value of the multigene signature was externally validated in two other sets from Gene Expression Omnibus (GSE14520 and GSE76427), and its role in recurrence prediction was also investigated. A total of 2000 DEGs were obtained, including 1542 upregulated genes and 458 downregulated genes. Subsequently, we constructed a 14-gene signature on the basis of 56 hub genes, which was a good predictor of overall survival. The prognostic signature could be replicated in GSE14520 and GSE76427. Moreover, the 14-gene signature could be applied for recurrence prediction in the training set and GSE14520. In summary, the 14-gene signature extracted from hub genes was involved in some of the HCC-related signalling pathways; it not only served as a predictive signature for HCC outcome but could also be used to predict HCC recurrence.