A Bioinformatics Approach to Investigating Significant Genes of Hepatocellular Carcinoma from Microarray Expression Data

Abstract

Hepatocellular carcinoma is one of most common cancers worldwide and also the top three most common causes of death from cancer in Taiwan. Hepatocellular carcinoma is often diagnosed after clinical deterioration and then the survival rate is low. Therefore, it is crucial to develop early diagnosis method in asymptomatic individuals. Accordingly, the study aimed to analyze, find and evaluate significant genes of hepatocellular carcinoma from microarray expression data by a systematic bioinformatics approach. The study clearly demonstrated the approach to investigate significant genes of hepatocellular carcinoma through significance analysis algorithm and various online databases. We firstly constructed a dataset from gene expression profiling of Affymetrix GeneChip on ArrayExpress database of functional genomics experiments. The dataset consisted of 455 samples in tumor and non-tumor tissues of hepatocellular carcinoma patients. Secondly, a significance analysis algorithm was employed to calculate candidate genes efficiently. Finally, we integrated the information obtained from COSMIC, catalogue of somatic mutations in cancer, and OncoDB.HCC, oncogenomic database of hepatocellular carcinoma, databases to evaluate the gene significance of hepatocellular carcinoma. From 22277 genes, 100 candidate genes were selected by the significance analysis algorithm. Further, we analyzed the studies of the candidate genes in COSMIC database and 98% of these candidates were related to cancer. In particular, 83% of these candidates were directly related to hepatocellular carcinoma. From the other hand, we analyzed the evidence of significant genes in OncoDB.HCC database and 41% of these candidates were supported by different types of evidence, including Stanford microarray, experiments and microarray/proteomic reports. The results showed that the proposal approach can effectively find and evaluate the gene significance of hepatocellular carcinoma from microarray expression data. Thus we suggested that the framework may be applied to other cancer studies based on microarray expression data and provide more possible clues to significant genes.