目的：通过生物信息学的方法探究糖尿病视网膜病变(DR)中免疫相关的关键基因以及免疫细胞的浸润情况。

方法：2022-09/10从GEO数据库获取基因芯片数据集，采用“limma”R包获得差异表达基因(DEGs)，并进行GO功能注释和KEGG通路富集分析，基于 CIBERSORT算法分析免疫细胞浸润情况。通过加权基因共表达网络分析(WGCNA)筛选与免疫相关基因模块中的DEGs，利用STRING在线数据库及Cytoscape软件构建蛋白质互作网络，利用MCODE以及cytoHubba插件进一步并筛选出关键基因。

结果：共筛选出上调差异基因1 426个，下调差异基因206个。原始B细胞、血浆细胞、记忆型CD4⁺T细胞、调节性T细胞(Tregs)、M0型巨噬细胞、M1型巨噬细胞以及中性粒细胞7种免疫细胞显著高表达(P<0.05)； NK cells activated这种免疫细胞低表达(P<0.05)。WGCNA分析后，与免疫最相关模块中差异基因820个，构建PPI网络后利用插件筛选出10个关键基因，利用各差异基因在PPI中的相关性程度进一步筛选出2个关键基因为DLGAP5与AURKB。

结论：利用生物信息学的方式筛选出DR患者中免疫细胞浸润情况以及与免疫相关的关键基因，可为DR的进一步研究与诊疗提供依据。

METHODS: Differential expression genes(DEGs)were obtained by “limma” R from Gene Expression Omnibus(GEO)data from September to October 2022, Gene ontology(GO)and Kyoto encyclopedia of genes and genomes(KEGG)were analyzed, and the infiltration of immune cell types in each sample was calculated based on CIBERSORT algorithm. Weighted gene co-expression network analysis(WGCNA)was used to screen for DEGs in immune-related gene modules. The protein-protein interaction(PPI)network was established by STRING online database and Cytoscape, and the hub genes were screened by MCODE and cytoHubba plug-ins.

RESULTS: The results showed that 1 426 up-regulated and 206 down-regulated differential genes were screened, where 7 immune cell types, including B cell naive, Plasma cells, CD4⁺T cells, T cells regulatory(Tregs), Macrophages M0, Macrophages M1 and Neutrophils were significantly overexpressed(P<0.05), while others were low expressed(P<0.05). After WGCNA, a total of 820 DEGs were found in the modules most related to immunity. After constructing the PPI network, 10 key genes were screened using plug-ins, and two key genes were further screened using the expression amount of each differential gene in PPI: DLGAP5 and AURKB.

CONCLUSION: This study used bioinformatics to screen the infiltration of immune cells and key genes related to immunity in patients with DR. These findings may provide evidences for future research, diagnosis, and treatment of DR.