Simulations of the EGFR - KRAS - MAPK Signalling Network in Colon Cancer. Virtual Mutations and Virtual Treatments with Inhibitors Have More Important Effects Than a 10 Times Range of Normal Parameters and Rates Fluctuations

摘要

The fragment of the signaling network we have considered was formally described as a sort of circuit diagram, a Molecular Interaction Map (MIM). We have mostly followed the syntactic rules proposed by Kurt W. Kohn [1, 10, 11]. In our MIM we drew 19 basic species. Our dynamic simulations involve 46 modified species and complexes, 50 forward reactions, 50 backward reactions, 17 catalytic activities. A significant amount of parameters concerning molecular concentrations, association rates, dissociation rates and turnover numbers, are known for this intensively studied neighborhood of the signaling network. In other cases, molecular, cellular and even clinical data generate additional indirect constraints. Some unknown parameters have been adjusted to satisfy these indirect constraints. In order to avoid hidden bugs in writing the software we have used two independent approaches: a) a more classic approach using Ordinary Differential Equations (ODEs); b) a stochastic simulation engine, written in Java, based on the Gillespie algorithm: we obtained overlapping results. For a quiescent and EGF stimulated network we have obtained a behavior in good agreement with what is experimentally known. We have introduced virtual mutations (excess of