Fig. 9
Methodology for the construction of an interaction map, derivation of the regulatory core, and prediction of molecular signatures using model simulations. a Data on molecular interactions are extracted from public databases and the literature to reconstruct the regulatory network underlying the cancer phenotype investigated. b Topological and non-topological network properties are used to identify important network motifs. A regulatory core is derived by merging the top-ranked motifs. c In silico simulations of the regulatory core help in the detection of molecular signatures that are then subject to experimental validation. d An algorithm for motif prioritization. Motifs can be prioritized based on the objective function represented in Eq. (1). This objective function contains parameters accounting for node properties, the expression profile of nodes in relevant cancer cell lines and their relatedness to disease pathways. Weights can be assigned by giving importance to a particular parameter in a user defined manner or iteratively to determine the Pareto sets of motifs. From these sets, top-ranked motifs can be identified based on user defined cutoff. e Logic-based representation of biochemical network. Left: A toy model of biological network consisting of four nodes (X1-4), and interactions among them regulate certain phenotype. Right: Derivation of Boolean functions (BF) and multi-valued logic functions (MF) for the toy model
