Abstract: Objective To explore the gene features and pathogenic mechanisms shared byhepatocellular carcinoma and diabetes mellitus using bioinformatics methods. Methods The hepatocellular carcinoma dataset (GSE121248) and diabetes mellitus dataset ( GSE29221) were downloaded from the GEO (Gene Expression Omnibus) database, and were analyzed by R. Venn diagrams were used to obtain the shared differentially expressed genes (DEGs). GO (gene ontology) and KEGG ( kyoto encyclopedia of genes and genomes) enrichment analyses were performed on DEGs, and Cytoscape software was used to obtain the key modules and core genes in the proteinprotein interaction (PPI) network. The network interaction analyses of genes, transcription factors and mi-RNAs were performed in NetworkAnalyst database. DGIdb database was used for gene-drug interaction analysis. The prognostic values of the core genes were analyzed by Kaplan-Meier Plotterdatabase. Results A total of 39 DEGs were screened out, which were significantly enriched inpathways such as extracellular matrix, positive regulation of insulin-like growth factor receptor signaling pathway, heparin binding, and P53 signaling pathway. 6 core genes ( THBS1, DCN, BGN, COL14 A1, LUM and PCOLC) were screened out, of which two core genes (DCN and THBS1) could interacted with some tumor therapeutic agents and one core gene (DCN) was associated with the prognosis of hepatocellular carcinoma patients. Conclusion DCN may be a potentialdrug therapeutic target for patients with both diabetes and hepatocellular carcinoma.

Key words: liver cancer, diabetes, bioinformatics, key gene, signal pathway, prognosis