目的：应用生物信息学方法分析视网膜母细胞瘤(retinoblastoma，RB)基因表达谱芯片，从转录水平探讨RB与正常视网膜组织的差异表达基因(differentially expressed genes，DEGs)的功能及相互作用。方法：通过公共基因表达数据库(gene expression omnibus，GEO)数据库中选取GSE97508和GSE110811两个芯片数据，共有34个RB组织和6个正常视网膜组织。利用GEO2R工具对和Draw Venn Diagrams软件，筛选出正常视网膜组织与RB组织间的DEGs。通过在线STRING检索工具(search tool for the retrieval of interacting genes，STRING)，对DEGs进行基因本体注释(gene ontology，GO)、KEGG富集通路分析、蛋白互作(protein-protein interaction，PPI)K分析。结果：在2个芯片中发现视网膜母细胞瘤共有20个DEGs，包括3个上调DEGs和17个下调DEGs，GO本体注释结果显示表达上调DEGs所富集的功能主要在细胞分裂、染色体浓缩、核分裂和DNA构象改变。下调DEGs主要富集在光传导、视觉感知、光感受器细胞修复、感光细胞内外节和调节视紫红质介导的信号通路等。KEGG通路显示上调DEGs没有显著的信号通路，下调DEGs参与光传导信号通路，其中包括CNGA1，CNGB1，RHO，SAG四个基因，通过PPI网络提示这4个基因相互联系，并发现与其他节点连接最紧密核心基因RHO。结论：利用生物信息学方法能有效对RB基因芯片数据进行挖掘，为进一步研究RB发生发展机制及寻找潜在的药物治疗靶点提供参考。

Objective: To find differentially expressed genes between retinoblastoma and normal retinal tissues by bioinformatics analysis, and to investigate their molecular function and interactions in the transcriptional level.Methods: The gene expression profile datasets GSE97508 and GSE110811, including 34 retinoblastoma (RB) tissues and 6 normal retinal tissues, were downloaded from gene expression omnibus (GEO). Differentially expressed genes (DEGs) between normal retinal tissues and RB tissues were identified by GEO2R tool and the Draw Venn Diagrams software. The gene ontology (GO) analysis, KEGG pathway and protein-protein interaction (PPI) were analyzed by STRING. Results: In two microarrays of retinoblastoma we found total 20 DEGs were identified, including 3 up-regulated and 17 down-regulated genes. The GO ontology annotation results showed that the enrichment functions of up-regulated genes were mainly in cell division, chromosome enrichment, nuclear division, and DNA conformation change. Down-regulated genes were mainly concentrated in light conduction,visual perception, photoreceptor cell repair, photoreceptor cell inner and outer segment, and regulation of rhodopsin mediated signaling pathway. The KEGG pathway showed that there was no significant signal pathway in which up-regulated genes up-regulated DEGs, and down-regulated genes were involved in the phototransduction signaling pathway, including four genes of CNGA1, CNGB1, RHO and SAG. PPI network suggested that these four genes were interlinked, and RHO was found to be the most closely connected core gene with other nodes.Conclusion: Bioinformatics can be used effectively to analyze RB microarray data to provide theoretical reference for further exploration of tumorigenesis mechanism and help search for potential drug therapeutic targets.