目的：通过生物信息学调研和分析，探讨对肝癌预后产生影响的独立危险因子及其与肝癌靶向药索拉非尼作用靶点的相互关联特征。方法：从TCGA数据库收集的248 例肝癌患者中选取79 例死亡患者的基因表达量、生存时间、生存状态数据，以中位生存时间360 d 为界将这些死亡患者分为2 组，用DEseq算法对这2 组患者的基因表达量数据进行差异基因筛选。不考虑其他混杂因素，仅考虑差异基因表达单个因素与患者生存时间相关，用Kaplan-Meier 法对差异基因进行单因素分析。将分析得到的差异基因表达量和患者生存时间、生存状态整合的矩阵用COX回归模型作生存分析，寻找对肝癌预后产生影响的独立危险因子。以STRING数据库为基础，将得到的差异基因与肝癌靶向药索拉非尼的作用靶点结合构建蛋白互联网络，探寻索拉菲尼作用靶点与差异基因的关系。结果：DEseq算法初筛得到52 个有可能对肝癌预后产生影响的差异基因。Kaplan-Meier 法分析得到26 个与患者生存时间相关的差异基因。COX分析最终确认3 个差异基因（SQSTM1、ANXA10 和STMN1）为影响肝癌预后的独立危险因子。其中ANXA10 为负向影响因素，而STMN1 与SQSTM1 为肝癌患者预后的正向影响因素。蛋白质相互作用网络显示，肝癌靶向药索拉非尼的作用靶点与多个差异基因密切相关。ANXA10 参与钙离子结合和钙依赖性磷脂结合，STMN1 和SQSTM1 在多种信号通路中起重要作用。结论：ANXA10、STMN1 和SQSTM1 可能是对肝癌预后产生影响的独立危险因子，可作为未来开发靶向药物的作用靶点或患者预后预测指标。

Objective: To explore independent risk factors affecting prognosis of the patients with liver cancer and their correlative features to action target points of targeting drug for liver cancer (sorafenib) via survey and analysis of bio-information. Methods: Data of gene expression, survival time and survival status of the 79 patients with liver cancer who finally died were selected from the 248 patients with liver cancer in the TCGA data base, the died patients were divided into 2 groups according to medium survival time of 360 days as a dividing line. Differential genes were screened out among data of the gene expression in patients of the 2 groups by DEseq algorithm. Without considering other mingled factors and only considering a single factor of differential gene expression and its correlation with survival time of the patients, uni-variate analysis of the differential genes was made by Kaplan-Meier assay.Survival analysis of a matrix integrated with the differential gene expression obtained by the analysis，survival time and survival status of the patients was made by COX regression model, in order to look for independent risk factors affecting prognosis of the patients with liver cancer. Based on STRING data base, a protein interconnected network was constructed with the obtained differential genes and action target points of sorafenib, targeting drug for liver cancer, to seek relationship of the action target points of sorafenib with the differential genes. Results: Fifty two differential genes that could affect prognosis of the patients with liver cancer were obtained by preliminary screening of DEseq algorithm. Twenty six differential genes relating to survival time of the patients were obtained by Kaplan-Meier assay. Three differential genes (SQSTM1, ANXA10 and STMN1) were finally confirmed as independent risk factors affecting prognosis of the patients with liver cancer by COX analysis. The protein-protein interaction network revealed that the action target points of sorafenib, targetin drug for liver cancer, were closely related to multiple differential genes. ANXA10 involved in combine of calcium ion and combination of calcium dependent phospholipids. STMN1 and SQSTM1 played a important role in multiple signaling pathways. Conclusion: ANXA10,STMN1 and SQSTM1 could be independent risk factors affecting prognosis of the patients with liver cancer, which might be used as action target points of targeting drug developed or prediction index of prognosis of the patients in future.

辽宁省博士科研启动基金资助项目（No.201501022）；军队医学科技青年培育计划资助项目（No.15QNP001）