The Differential Rank Conservation Algorithm (DIRAC) Reveals Deregulation of Central Metabolic Pathways in Hepatocellular Carcinoma

Abstract

Hepatocellular carcinoma (HCC) is a deadly disease without existing cure or effective treatment. If the genetics and metabolics of this form of liver cancer would be better understood the chance of developing effective treatments would likely increase. HCC cells are different from normal liver cells. When it comes to gene expression the traditional way is to define these differences in terms of up- or downregulation of genes or metabolic subnetworks. The purpose of this report however is to present an alternative approach regarding the definition of what makes HCC cells different from matched liver cells. This approach does not primarily focus on between-phenotype differential gene expression but rather on metabolic network regulation and can be viewed in terms of metabolic network entropy. This alternative approach is made possible through implementation of the so called Di_erential Rank Conservation algorithm (DIRAC). In addition traditional differential expression analysis and detection of differentially expressed metabolites by the reporter metabolites algorithm was performed. The basis of the analysis was RNA-Seq data from 163 HCC patients downloaded from The Cancer Genome Atlas database. The results indicate fundamental alterations in network regulation of central metabolic pathways in HCC. For example the TCA cycle, the elec-tron transport chain and fatty acid metabolism show general apparent dysfunction with network deregulation and concomittant average down-regulation of gene expression in HCC. In contrast fatty acid biosynthesis seems to be under tight regulatory control with average upregulation of gene expression in HCC. The DIRAC algorithm also revealed large scale average metabolic network disorganization in HCC, a phenomena referred to as global deregulation. In addition six other cancers were also analysed in terms of their global regulation of their metabolic networks. This analysis revealed a correlation between the degree of global deregulation and malignancy of the cancer. In conclusion the DIRAC algorithm o_ers an alternative view on between-phenotype differences of gene expression which reveales important alterations of the metabolism in HCC.