目的：运用生物信息学方法结合基因表达数据库（Gene Expression Omnibus，GEO）筛选参与胃癌发生发展的关键基 因，以获得用于胃癌诊断、治疗靶标选择及预后判断相关分子标志物。方法： 从GEO数据库中下载与胃癌（gastric cancer，GC）相 关芯片数据集，筛选差异表达基因（differentially expressed gene，DEG）， 对DEG 做功能富集分析，构建蛋白互作网络（proteinprotein interaction network，PPI），筛选关键基因（key gene），进而构建共表达网络、生存曲线以及层次聚类分析。结果：共筛选出 261个GC相关DEG，通过分析获得14个关键基因，分别为PLOD1、PLOD3、COL1A1、COL1A2、COL2A1、COL3A1、COL4A1、 COL4A2、COL8A1、COL12A1、COL15A1、ITGA2、LUM、SERPINH1。关键基因主要参与胶原纤维的组织、细胞外基质的组织、 细 胞外结构的组织、皮肤形态发生、胶原的合成及血管的发育等生物学过程。生存曲线分析表明，基因COL3A1（P=0.0241）表达的 改变显著降低了胃癌患者整体生存率；基因ITGA2（P=0.0679）的表达改变也显示与GC患者的无病生存率降低有关。与正常胃 组织相比，层次聚类分析表明，GC组织中基因PLOD1、PLOD3、COL3A1、ITGA2、COL1A2、COL1A1、COL4A1、LUM、COL12A1、 SERPINH1、COL8A1表达上调。结论：经筛选获得的关键基因可作为潜在分子标志物用于GC的早期诊断、治疗靶点选择和预 后判断，并为后续的研究提供参考。

Objective: Bioinformatics combined with Gene Expression Omnibus (GEO) was used to screen key genes involved in the development of gastric cancer in order to obtain molecular markers for diagnosis, target selection and prognosis prediction of gastric cancer. Methods:Thechipdatasetsrelatedtogastriccancer(GC)fromtheGEOdatabase were downloaded, and differentially expressed genes (DEG) were screened. Functional enrichment analysis on DEG was performed, and protein-protein interaction network (PPI) was constructed to screenkeygenes.Then,co-expressionnetworkswerefurtherconstructed,andsurvival curves were drawn and hierarchical clustering analysis was performed. Results: A total of 261 GC-related DEGs were selected, and 14 key genes were obtained through analysis, which were PLOD1, PLOD3, COL1A1, COL1A2, COL2A1, COL3A1, COL4A1, COL4A2, COL8A1, COL12A1, COL15A1, ITGA2, LUM and SERPINH1. Key genes are mainly involved in biological processes such as generation of collagen fiber tissues, extracellular matrix tissues, extracellular structure tissues, skin morphogenesis, collagen biosynthesis and vascular development. Survival curve analysis showed that the change in the expression of COL3A1 (P=0.0241) significantly reduced the overall survival rate of patients with gastric cancer; the change in the expression of ITGA2 (P=0.0679) also showed a correlation with the reduction of diseasefree survival in gastric cancer patients. Compared with normal gastric tissues, hierarchical cluster analysis showed that the expressions of genes PLOD1, PLOD3, COL3A1, ITGA2, COL1A2, COL1A1, COL4A1, LUM, COL12A1, SERPINH1 and COL8A1 in GC tissues were up-regulated. Conclusion: The key genes obtained after screening can be used as potential molecular markers for early diagnosis, treatment target selection and prognosis judgment of gastric cancer, which provide reference for subsequent research.

山西省自然科学基金资助项目（No.201901D111412，201901D111232）；山西省研究生教育创新项目资助（No.2020SY197）