Simulation of signal transduction pathways

Signal transduction is about how molecular mechanisms in cells are used to receive, process and respond to signals from outside and inside of the cell. The term signal is defined in a dual way in this context: being the molecular state of a molecule, e.g. a protein, and at the same time an abstract logical state (active/inactive). This thesis focuses on intracellular signalling: i.e. receiving some kind of stimuli (e.g. hormones) from outside, integrating this information with the state of the cell, and passing this information into the nucleus to regulate gene expression by transcription factors as the response to the stimulus. In recent years a huge amount of data has been accumulated in various biological databases. These data are used for quantitative analysis of small, well-defined signal networks. The dissertation follows the new approach of qualitative simulation of signal transduction on whole-cell scale. In collaboration with the company Biobase/Wolfenbüttel, Germany the signal transduction database TRANSPATH Professional is used to gain access to qualitative, manually annotated textual information from the scientific literature. For qualitative simulation of signal transduction pathways a model of interacting signal molecules is established, similar to the information structure of TRANSPATH . Signalling behaviour is modelled in a process-oriented, as well as in an object-oriented way. Signal molecules are represented as entities that have variables to keep their conformation status or biochemical alterations, and as parameterised procedures to communicate with each other. In a first approach the established symbolic signal model was implemented in the π-calculus. A simulation software, based on this theory and called PsiFCP, is used to simulate the well characterised Egf-Mapk signal pathway. As intermediate result of the simulation with the PsiFCP system, software requirements for simulation of signal transduction networks are formulated. In a second attempt the object-oriented signal model was implemented in an agentbased simulation environment, called Swarm system. An extended Egf-Mapk signal pathway is reimplemented in the programming language Objective C to feed the Swarm simulation system with local, binary interaction data of TRANSPATH 4.2. Swarm then dynamically processes the signal flow using a stochastic scheduling mechanism, which was developed to this end. Simulated models can be analysed by inspecting signal flow and scheduling traces, and the shortest path between two signal molecules can be computed. As final result, shortest paths are computed between varying molecules of the manually imported Egf-Mapk pathway.