Figure 1
The process for genome-scale model reconstruction. (1) Metabolic properties including associations between genes, enzymes and reactions were extracted from the genomic metabolic databases, stored and curated in an Excel spread-sheet; (2) a 2D reaction centric SBML representation was generated using an in house java application. (3) The stoichiometric matrix, as well as reversibility constraints and the objective function were extracted from the SBML; (4) the relevant linear programming problems were solved using the COBRA toolbox [37]. (5) The model was refined in an iterative process, accessing the best available information in the literature and online data sources to achieve metabolic functionality.