Objective: To explore changes for circRNAs expression profiles in gastric cancer, as well as the changes of competitive endogenous RNA network regulation of circRNAs-miRNAs-mRNAs. Methods: GSE78092, GPL19978 (a combination of GSE83521 and GSE89143), GSE100170, and STAD(unavailable for the public) were selected from the gene expression omnibus（GEO）database. The total of circRNAs were collected from these data. Then through Starbase, TargetScan and other bioinformatics analysis was performed. Results: The total of 734 differentially expressed circRNAs were obtained from gastric cancer tissues and paracancerous tissues, and 17 circRNAs were screened according to P<0.05, log (fold change) >1.5. Then the interaction analysis of circRNAs and microRNAs (miRNAs) using Starbase database was performed again, a total of 354 target miRNAs were found, which were verified by GEO database (GSE23739), 46 miRNAs were finally obtained, and the targeted mRNAs were analyzed using TargetScan database, the kyoto encyclopedia of genes and genomes(KEGG) analysis and gene ontology (GO) analysis showed that target mRNA can promote or inhibit the occurrence and development of gastric cancer mainly in the aspects of transcriptional regulation, RNA metabolism, intracellular signaling cascade and protein localization. Conclusion: The 17 differentially expressed circRNAs formed a competitive endogenous RNA network of circRNAs-miRNAs-mRNAs, which indicated that circRNAs-miRNAs-mRNAs could regulate the occurrence and development of gastric cancer by regulating transcription and RNA metabolism, and could be used as a predictor of gastric cancer and provide a reference for clinical diagnosis and treatment of gastric cancer.

［Key words］circRNAs；gastric cancer；circRNAs expression profiles；biomarker