Kanserde metabolik deregülasyonun değerlendirilmesi için transkriptom verilerinin analizi

The great potential of cancer metabolism in tumor development as well as treatment has been appreciated in recent years. In this study, an extensive list of enzymes relevant to central cellular metabolism and metabolite transporters were manually curated. A pan cancer differential expression of these genes between tumor and their normal counterpart was analyzed in an attempt to better understand tumor bioenergetics, catabolism and anabolism in general. Major deregulation of these genes, particularly the enzymes involved in NADPH metabolism was observed in Liver Hepatocellular Carcinoma (LIHC). LIHC is an extremely aggressive treatment resistant disease with poor prognosis and overall survival (OS), probably due to the ability of the organ to detoxify chemotherapy drugs. AKR1B10 is an NADPH utilizing enzyme whose expression was increased from a very early stage in most LIHC patients and was associated with low OS. However, mechanistic underpinnings of this association is unknown. To address this and the context in which it occurs we hypothesized a potential connection of AKR1B10 with the pentose phosphate pathway (PPP), a glucose metabolizing pathway that generates NADPH. Using various bioinformatics and computational approaches, we found that AKR1B10 works in concert with PPP and a number of other detoxifying enzymes to enhance tumor aggressiveness. In addition, we report the use of latest tools in machine learning to build multi-gene signature prognostic models for highly significant LIHC patient stratification.