Identification of recurrence-related genes by integrating microRNA and gene expression profiling of gastric cancer.

We previously analyzed the microRNA (miRNA) expression pattern in gastric cancer with and without recurrence and obtained 17 differentially expressed miRNAs with potential to predict recurrence risk for GC patients. In the present study, we aimed to investigate recurrence-related genes which may be regulated by the differentially expressed miRNAs identified in our prior research. Three different miRNA target gene databases (miRanda, TargetScan and PicTar) were used for searching the potential genes regulated by miRNAs. A combination was performed between miRNA target genes and recurrence-related gene expression profiling. Three bioinformatics algorithms (PAM, SVM and RF) were used to feature recurrence-related gene selection. In addition, we validated the expression levels of the genes in GC patients using real-time PCR. A total of 3,263 genes were identified as potential targets of 17 miRNAs. We identified 2,736 differential expressed genes using the SAM method based on 22K oligo microarray data which included 7 recurrence and 4 without recurrence GC samples. Combining the target genes regulated by miRNAs and the differentially expressed genes between recurrence and non-recurrence groups, we identified 228 differential genes for further study. Finally, we identified HNRPA0 and PRDM4 as risk biomarkers of GC patients, which were regulated by hsa-miR‑194 and hsa-miR-373, respectively. Our data indicated that HNRPA0 and PRDM4 may be involved in the recurrence process of GC and have potential to act as new prognostic biomarkers in predicting recurrence risk for gastric cancer patients.