From reaction networks to information flow--using modular response analysis to track information in signaling networks

Abstract

Even if the biochemical details of signaling networks are known, it is often hard to track how information flows through the network. In combination with experimental techniques, modular response analysis has proven useful in analyzing the quantitative information transfer in signal transduction networks. The sensitivity of a target (e.g., transcription factor, protein) to an upstream stimulus (e.g., growth factor) can be determined by a so-called response coefficient. We have used this methodology to analyze how information flows in networks where the details of the mechanisms in the networks are known, but parameters are lacking. Using a Monte Carlo approach, we apply this method to track the routes of information flow. More specifically, we determine whether a given species has no, positive or negative influence on any other species in the network. Surprisingly, one can uniquely determine whether a molecule activates or inhibits another one in more than 99% of the interactions solely from the topology of the reaction network. To exemplify the methodology, we briefly discuss three signaling networks of different complexity: (i) a Wnt signaling pathway model with 15 species, (ii) a MAPK signaling pathway model with 200 species, and (iii) a large-scale signaling network of the entire cell with over 6000 species.