Functional analysis of the genes in the glucose metabolic network by a system based modular approach

Abstract

In this study, all proteins associated with the Process Gene Ontology (GO) term “Glucose Metabolic Process” were collected and a network of all reported interactions among these entities was constructed with a stringent confidence threshold. Two networks were constructed at the beginning of the study; one general network consisting of a global interaction network of yeast obtained from the STRING database (Global Network, GN) and another network that was constructed by including only the interactions for the nodes of the Glucose Metabolic Process Gene Ontology Term (Glucose Metabolic Network, GMN). These two networks were then compared and the GMN was determined to be a scale free biological network. The parameters used in module identification were optimized in order to increase module specificity. Modular structures of GN and GMN were investigated by using Molecular Complex Detection (MCODE) plug-in of Cytoscape. The modules of GN had more members and these modules were associated with very general GO terms but the modules that were identified in GMN were more compact and functionally more distinctive with significant gene ontological associations. 65 modules were identified in GMN in the present study and eight five membered modules were selected and analyzed. These modules were associated with a specific function in the cell, either as a part of a complex or as a part of a signaling cascade. The extension of the modules provided additional nodes that were associated with the same function or other functions. All associations were verified from the available literature on the investigated pathways and complexes. The functional robustness of the method in identification of module members associated with the same cellular process suggests that this modular approach can be used to assign probable functions to the unannotated or uncharacterized loner genes present, or to assign novel additional functions to the proteins with already known functional associations. As a result of this investigation, YGR067c was potentially related to “Ubiquitin Dependent Protein Catabolic Process”, Roy1p was potentially related to “Interphase” and Nnk1p was potentially related to “Proteolysis” GO Process Terms.