Integration and analysis of protein-protein and signalling networks to create graph dynamical systems

Understanding networks of biological interactions is essential to all life sciences. Nowadays, a large amount of biological information is archived in diverse resources that often contain complementary information. Scientists have to search various data repositories in order to reconstruct, understand and speculate on the steps that lead to specific phenomena. The amount of digitalised interaction data has grown to such an extent that it is now advisable to integrate different kinds of data through computational means. Bioinformatics can offer tools and strategies to manipulate such information automatically. In order to address the need of integrating and retrieving structured protein interaction data, I implemented a new resource called mentha, which merges the data archived in different repositories in order to assemble large networks of protein interactions. This resource has low data redundancy and it assigns a reliability score to each interaction. mentha can be seen as a workbench to assemble custom protein-protein interaction networks. Furthermore, I contributed to the development of a new resource that archives manually curated signalling information. I investigated the possibility of exploiting the information contained in mentha to assemble signalling networks. I analysed strategies to automatically link proteins of interest with other relevant proteins in order to create a backbone to build graph dynamical systems that can represent biological signalling networks. I conclude that manually curated signalling information can be integrated with information inferred from interactomes where signal direction can be predicted using orientation algorithms, while signal effect can be derived from RNA interference screenings. Finally, I analysed the resulting biological models as graph dynamical systems. I transformed the topology of a signed oriented biological network into transition boolean functions and, using boolean formalism, I analysed dynamical networks to simulate biological behaviours.