Reconstruction of Biological Networks for Integrative Analysis

Biological systems can be very complex and consist of several thousand components that interact with each other in the cell. One of the goals of systems biology is to study biological systems from a systemic viewpoint in order to get an increased understanding of the behavior of the cell. Biological network reconstructions are important tools in systems biology in order to model the behavior of different biological systems. The biological networks can also be used as a scaffold for integrative analysis where high-throughput data from different conditions or different strains are integrated into the biological network to reduce the dimension of the data and to group the response between conditions or strains into biological pathways or key metabolites etc. The biological interpretation and discovery using integrative analysis can be facilitated by constructing more comprehensive and diverse biological networks.In this thesis I expanded current biological network reconstructions for the yeast Saccharomyces cereveisae in three steps and used them as a scaffold for biological interpretation and discovery. First I constructed an up-to-date yeast genome-scale metabolic model. The model is a comprehensive description of yeast metabolism and contains more genes, reactions and metabolites than previous models. The model performs well in simulating the metabolism under different conditions. Second, I studied the transcriptional regulatory network of yeast in terms of topology and structure of the network and compared it to transcriptional regulation in E. coli, human and mouse. I also used high-throughput data from many different conditions to study the condition-dependent response of the yeast transcriptional regulatory network. Third, I was involved in reconstruction of models of the protein secretion machinery in S. cerevisiae and for the high protein producer Aspergillus oryzae, describing protein folding, post-translational modifications and protein transport etc. High-throughput data from several different strains producing α-amylase were integrated into the models in order to get an insight in the mechanisms and bottlenecks of protein secretion in these organisms.The biological networks presented here were also used for data integration and the results and interpretation of the cellular behavior under different conditions can give us a deeper understanding and insight in for example condition-specific transcriptional regulation and protein production.