Figure 1

A schematic figure explaining the metabolic coherence ( MC ), inconsistency I, and methodological approach behind the comparison of MC and I. (a) A threshold t is applied onto the data x, depicted here as an overlay histogram of the raw log-signal intensities of the adenoma data, in order to obtain a binary gene/reaction (nodes) presence (green) and absence (red) pattern p. (b) Effective gene subnetworks are constructed from the present genes in p and the overall static gene-network representation of human metabolism (step 1). The ratio (coherence C) of connected genes (gray nodes) to isolated genes (light gray nodes) is determined (step 2). Sampling the overall static network with randomly generated presence patterns provides a distribution of null hypothesis coherence values C’ (step 3). The coherence C is transformed into the z-score MC (metabolic coherence) using C’ (step 4). (c) GIMME (Gene Inactivity Moderated by Metabolism and Expression) computes suitable flux distributions by simultaneously asserting flux through a specified objective (v^obj; in fact a certain level l of the theoretically achievable maximum$v_{\mathit{\text{max}}}^{\mathit{\text{obj}}}$ determined by FBA) and minimizing flux through absent reactions. (d) The flow chart depicts the necessary steps in setting up a comparative analysis of MC and I.