Applications of convex optimization in metabolic network analysis

Digital content

Author/Creator: Tefagh, Mojtaba, author.
Contributor: Boyd, Stephen P. degree supervisor. Thesis advisor; Tse, David degree committee member. Thesis advisor; Ye, Yinyu degree committee member. Thesis advisor; Stanford University. Department of Electrical Engineering.

Summary: Twenty years ago, the first genome-scale metabolic network reconstruction of the cellular metabolism of an organism was published, shortly after the first genome was sequenced. From that time on, the ever-increasing advances in the high-throughput omics technologies have allowed for the comprehensive reconstructions of exponentially growing sizes. However, the vast amount of data can be a two-edged sword which makes many essential tasks computationally intractable. To overcome the demands of systems biology, even while they are outpacing Moore's law, faster computational techniques are needed to enable the current methods to scale up to match the progress of data generation in a prospective manner. In this dissertation, we go over several different areas of systems biology from flux coupling analysis to context-specific reconstruction and propose efficient computational methods for several tasks separately. Then we work towards a more holistic approach and discuss the idea of a canonical metabolic network reduction to reduce the number of reactions for any general task. In analogy to the concept of lossless compression in information theory, we will show that the well-known emergent redundancies of flux distributions are the key concept. Additionally, we will derive the minimum reduced metabolic network and prove a converse that any further reduction loses some information on the elementary modes.