Student Ph.D. Defense: Pathway Analysis of Metabolic Networks using Graph Theoretical Approaches

Abstract

Cellular pathways defining biochemical transformational routes are often utilized as engineering targets to achieve industrial-scale production of commercially useful biomolecules including polyesters, building blocks for polymers, biofuels, and therapeutics derived from isoprenoids, polyketides, and non-ribosomal peptides. Identifying target pathways can be expedited using computational tools, leading to reduced development cost, time, and effort, and enabling new discoveries with potential positive impact on human health and the environment.

This thesis addresses three cellular pathway identification problems within metabolic networks. In the first problem, we identify all stoichiometrically balanced, thermodynamically feasible and genetically independent pathways, known as Elementary Flux Modes (EFMs), that can be used to express flux distributions and characterize cellular function. We develop an algorithm, gEFM, that incorporates structural information of the underlying network to enumerate all EFMs. The results show that gEFM is competitive with state-of-the art EFM computation techniques for several test cases, but less so for networks with larger number of EFMs. In the second and third problems, we identify individual target pathways with pre- specified characteristics. We develop an algorithm, PreProPath, for identifying a target pathway for up-regulation such that the path is predictable in behavior, exhibiting small flux ranges, and profitable, containing the least restrictive flux-limiting reaction in the network. The results show that PreProPath can successfully identify high ethanol production pathways across multiple growth rates, and for succinate production in Escherichia coli (E. coli) as published in the literature. We also develop an algorithm, Dominant-Edge Pathway, that identifies thermodynamically-limiting reactions along a pathway within the network from a given source metabolite to the desired destination. The algorithm is used to identify thermodynamically-limiting pathways in Zymomonas mobilis (Z. mobilis), E. coli and rat liver cell.

The novelty of this thesis is in utilizing graph-based methods to enumerate EFMs and to efficiently explore the pathway design space. Overall, the thesis advances the state-of-the-art techniques for metabolic pathway analysis.