Year: 2026
Type: dissertation
Abstract: Cellular metabolism represents the fundamental processes of life, enabling organisms to grow, adapt, and produce valuable compounds. However, its complexity poses a major challenge for both systems biology and metabolic engineering. Metabolism cannot be understood as an isolated process; instead, it is an interconnected network deeply embedded within cellular physiology. Metabolic reactions are regulated not only by stoichiometry and thermodynamics but also by processes such as gene expression, transcriptional and translational regulation, enzyme availability, signaling pathways, and epigenetic modifications. These layers of control determine whether a given pathway is active, how metabolic fluxes are distributed, and how cells adapt to the environment. Thus, metabolism is inherently coupled to other cellular processes, and any attempt to model it in isolation overlooks critical regulatory and dynamic influences. For this reason, traditional approaches—such as constraint-based or kinetic models—seem insufficient on their own. Constraint-based models ignore multiple constraints, while kinetic models need extensive data on kinetic parameters. Both frameworks miss the interconnection between metabolism and the broader cellular context. To overcome this, there is a pressing need for integrative modeling approaches capable of combining multiple biological data types and linking different scales of analysis. This thesis addresses the limitations of traditional modeling approaches by bridging the gap between constraint-based and kinetic models to produce more accurate and physiologically consistent predictions. Beyond this integration, it also introduces Flexible Nets (FNs) as a novel modeling formalism that unifies multiple data types—kinetic, proteomic, and environmental—and genome-scale metabolic modeling. First, the integration of kinetic constraints into an Escherichia coli genome-scale model was investigated. By embedding kinetic bounds into the genome-scale model, the enriched model reduced the number of dormant reactions and provided more realistic simulations of pathway usage. Furthermore, we solved a bifurcation problem in the genome-scale model, in which flux was previously directed entirely either towards growth or towards product—citramalate—formation. By incorporating kinetic bounds to the genome-scale model, we established a trade-off that allows simultaneous, non-zero fluxes through both the growth and product formation. Second, the thesis advances the integration of kinetic and proteomic data in a Bacillus subtilis model, marking one of the first applications of FNs with enzymatic constraints. The combined framework successfully narrowed the overestimation of scyllo-inositol production observed in unconstrained models and highlighted condition-specific bottlenecks such as enzyme availability and nutrient uptake. Finally, the work expanded the scope of metabolic modeling to multi-scale systems by coupling intracellular networks with extracellular bioreactor dynamics. Two case studies were explored: citramalate production in E. coli and monoclonal antibody (IgG) production in Chinese Hamster Ovary (CHO) cells. In both scenarios, FN-based models reproduced continuous culture dynamics, guided the optimization of operational conditions, and in the second case study, nutrient minimization of the medium was performed in order to find an alternative to commercial media. Overall, this thesis advances metabolic modeling by bridging constraint-based and kinetic approaches and by introducing FNs as a powerful new formalism. These contributions enhance the realism of metabolic predictions while extending their applicability to multi-scale bioprocess design.
Source: Zaguan (Universidad de Zaragoza)
Institution: Universidad de Zaragoza
Language: English
Subfield: Molecular Biology
Domain: Life Sciences
Open Access status: green
OpenAlex ID: W7225470562
