| Signaling pathways in mammalian cells can be combined to form networks. These networks can be represented as digraphs where biomolecules are nodes and interactions are represented as links. Graph theory analysis can be applied to understand the topology of such networks. A mammalian neuronal regulatory cellular signaling network was extracted from biomedical literature into a qualitative abstract template. The network was found to be scale-free and small-world, and was analyzed using original algorithms to identify network motifs in subnetworks. It was found that the network is enriched in bifan and bi-parallel regulatory motifs, while pathways are clustered underneath the most influential ligands. Negative feedback loops are mostly located close to the membrane, whereas positive feedback loops are enriched in general. Methods to measure the dynamical stability of such network map using random Boolean simulations were developed. These methods were applied to analyze the signaling network developed in this project, and gene regulatory networks of yeast and bacteria. The analysis showed that distribution of signs associated with links (positive/negative) may contribute to dynamical stability. Additionally, analysis of growing and adapting scale-free, exponential, and duplication-divergence artificial networks was implemented. Artificial networks, after growth and adaptation, were compared to the neuronal signaling network. Qualitative similarities were observed. Software tools to build, analyze, and visualize cell signaling networks were also developed: McSEDER, a search engine and data-mining tool, SAVI, a desktop software, and PathwayGenerator a web-based information system. The software systems are useful resources for biomedical researchers studying cell signaling. |
