On metabolic networks and multi-omics integration

Doctoral thesis, 2013

Cellular metabolism is a highly complex chemical system, involving thousands of interacting metabolites and reactions. The traditional approach to understanding metabolism has been that of reductionism; by isolating and carefully measuring the involved components, the goal has been to understand the whole as the sum of its parts. This reductionist approach has successfully identified most of the components of metabolism but, unfortunately, it fails to capture the long-range and complex interactions that are essential for the functionality. Systems biology is an emerging research field which uses high-throughput data generation and mathematical modelling in order to apply a holistic, or network-centric, view on metabolism. One type of modelling framework, which is in line with this thinking, is genome-scale metabolic modelling. These models, called GEMs, represent very valuable resources, but their applications have been limited due to the large manual effort required to reconstruct them. In this project, we have developed algorithms and software for streamlining the reconstruction process, as well as for novel applications of GEMs. More specifically, we here present: the RAVEN Toolbox, a software suite for automated reconstruction and quality control; the INIT algorithm, an algorithm for inferring GEMs for human cell types; an algorithm which integrates fluxomics and transcriptomics data in order to identify transcriptionally controlled metabolic reactions. The methods and software were used in a number of case studies to address real biological questions. These studies were: 1) Metabolic engineering of Saccharomyces cerevisiae for succinic acid overproduction. The predictions from the modelling were successfully validated experimentally. 2) Study of metabolic regulation in S. cerevisiae. This led to the identification of a small number of transcription factors and enzymes which were predicted to be controlling central parts of metabolism. 3). Penicillin production in Penicillium chrysogenum. This led to the reconstruction of the first GEM for P. chrysogenum, an important resource in itself, and to identification of metabolic engineering targets for more efficient production of penicillin. 4) Human cancer metabolism. This led to the identification of metabolic subnetworks which were predicted to be significantly more active in cancers, and to identification of potential drug targets for treatment. 5) Lipid metabolism in obesity. This led to new insights into the large-scale metabolic rearrangements associated with obesity, and to identification of possible therapeutic strategies. 6) Metabolism in non-alcoholic fatty liver disease. This led to the identification of serine deficiency as a central aspect of the disease, and to proposed therapeutic strategies for remedying it. The work put forward in this thesis has resulted in improvements on several important aspects of genome-scale metabolic modelling, and it has shown how the framework can be applied to gain novel biological insights. As such, it can contribute to further increase the role of the framework in modelling of human health and disease.