Abstract: Objective To construct a prognostic circRNA-miRNA-mRNA ceRNA network ingastric cancer ( GC) and explore the diagnostic and therapeutic targets. Methods Differentiallyexpressed circular RNAs (circRNA), related microRNAs (miRNA) and mRNAs in GC were obtained from GEO and TCGA databases. Cytoscape was utilized to construct the circRNA-miRNA-mRNA ceRNA network, to understand the biological functions of differentially expressed mRNAs and identifying the hub genes. Kaplan-Meier survival analysis was used to verify the prognostic values. Results The ceRNA network was consisted of 2 circRNAs, 5 miRNAs and 349 mRNAs thatwere differentially expressed. Functional enrichment analysis revealed that mRNAs were significantly enriched in terms and pathways like “ muscle system process”, “ metal ion transmembrane transporter activity”, “synaptic membrane” and “calcium signaling pathway” . A PPI network containing 125 nodes and 165 edges was constructed and 10 hub genes were identified. Survival analysis implied that low ADAR1D or PTGFR expression levels might lead to poor prognosis of GC patients. Conclusion The circRNA-miRNA-mRNA ceRNA network involved in prognosis of GC issuccessfully constructed. Two axes, the hsa _ circ _ 0001190 / hsa-miR-7-5p / ADRA1D and the hsa_ circ_ 0036287 / hsa-miR-3127-5p / PTGFR, are identified with potential prognostic values in GC, providing a new target for diagnosis andtreatment of GC.

Key words: circular RNA, gastric cancer, ceRNA, mRNA, prognosis