[Bioinformatics on vascular invasion markers in hepatocellular carcinoma via Big-Data analysis]

Objective: To investigate the biomarkers in hepatocellular carcinoma and their prognostic value via GEO (Gene Expression Omnibus) and TCGA (The Cancer Genome Atlas) database. Methods: Datasets of hepatocellular carcinoma were downloaded from GEO (GSE67140) and TCGA. MicroRNA in SNU423, SNU449, HepG2, Hep3B, SNU398 cell lines which had low or high invasion capabilities were investigated and verified, in 81 patients with and 91 without vascular invasion hepatocellular carcinoma. The prognostic value of these microRNAs were studied via TCGA database,obtained from 362 patients with hepatocellular carcinoma, through Kaplan-Meier and Multivariate Cox proportional hazard analysis. Target genes were analyzed by GO and KEGG. Results: Expressions of hsa-mir-1180, hsa-mir-149, hsa-mir-744 and hsa-mir-940 were all up regulated in high invasion capable cell lines (SNU423, SNU449) and vascular invasion patients with hepatocellular carcinoma (logFC>1, P<0.05). Results from the Survival analysis showed that hsa-mir-1180 (HR=1.623, 95% CI: 1.114-2.365, P=0.012), hsa-mir-149 (HR=2.400, 95% CI: 1.639-3.514) and hsa-mir-940 (HR=1.704, 95%CI: 1.188-2.443, P=0.004) were independent risk factors on the prognosis of patients with hepatocellular carcinoma (P<0.05). The mechanism might be related to factors as immune response, focal adhesion and adherence junction signaling pathways. Conclusion: With TCGA and GEO data mining, we found that hsa-mir-1180, hsa-mir-149, hsa-mir-744 and hsa-mir-940 were all highly related to the prognosis of hepatocellular carcinoma, that enabled it to be used to further study the biomarkers related to the prognosis of hepatocellular carcinoma.