Fig. 3

Sample trajectories of the integrative E. coli network. Initial perturbations differ in size, q = 1 − p = 0.01 (a, b) and q = 0.03 (c, d), as well as in localization: perturbations are initiated in the gene regulatory domain (a, c) and the metabolic domain (b, d). In each panel, the vertex states are plotted against the update steps according to Eqs. (1) and (2). The vertices are grouped according to the domain affiliation, starting from gene regulatory domain (top, green) via protein-interface domain (middle, blue) towards metabolic domain (bottom, red). Within each domain they are arranged with respect to the update step, in which they were affected in the respective run (vertex ordering within the domains is thus different from run to run). Affected vertices are shown in the color of their domain, unaffected vertices are shown last for each run and in white. The horizontal dashed lines denote the domain boundaries and the black square indicates the domain with the largest number of newly affected vertices in each step. Connections between black squares are only shown to guide the eye