Resumen
Currently, there is no effective method to detect the prognosis for hepatocellular carcinoma (HCC). This study used bioinformatics techniques to determine HCC molecular subtypes and prognosis-related biomarkers. A total of 3330 intersectional differentially expressed genes (DEGs) with the same differential direction in four datasets were identified by differential expression analysis. Intersectional DEGs were involved in the cell cycle, FOXO signaling pathway, and complement and coagulation cascades. Then, two subtypes were identified using a non-negative matrix decomposition method. Subtype C2 displayed better overall survival than subtype C1. Moreover, 217 prognostic related-genes were identified using the Cox regression and Kaplan-Meier curves. The area under the curve >0.80 of prognostic relate-genes were selected to construct random survival forest and the least absolute shrinkage and selection operator model and obtained seven feature genes (SORBS2, DHRS1, SLC16A2, RCL1, IGFALS, GNA14 and FANCI). Risk score model and recurrence model were constructed based on feature genes, and FANCI was inferred as a key gene by univariate Cox model. High expression of FANCI was mainly involved in cell cycle, DNA replication and mismatch repair. Interestingly, Single Sample Gene Set Enrichment Analysis was used to quantify immune infiltration and showed that Th2 cells and T helper cells were significantly up regulated in HCC compared to controls. Furthermore, we found the presence of two mutation sites as well as methylation modifications occurred in FANCI. Overall, we identified two types of HCC and identified that FANCI will serve as a potential biomarker for HCC prognosis and be important to the diagnosis and treatment of HCC.