Project Title
Computational Tools for Metabolic Engineering using Genome-scale Integrated Models
Project Type
Nacional / Public
Funding Body
Funding Program
  • CEB: 0,00
  • Total: 111 940,00
Universidade do Minho
External link

Principal Investigator

Team Members - CEB


In the last few years, the field of Metabolic Engineering (ME) has grown consistently, driven by the increased adoption of Biotechnology in the production of important chemicals, mainly in the pharmaceutical and food industries [1]. ME deals with the rational design of improved organisms by genetic modifications of existing strains, which are able to achieve higher productivities in a desired production process [2]. A number of approaches to ME have been proposed, relying on mathematical models of the metabolism combined with computational algorithms and experimental data to identify suitable targets for genetic engineering. Some of these techniques, like Metabolic Control Analysis, use dynamical models, while others like Metabolic Flux Analysis or Flux Balance Analysis, take steady-state models to simulate the phenotype of microorganisms under different environmental and genetic conditions. These efforts led to some successful examples [3], but the methodologies are still limited in effectively achieving the rational design of improved strains. The proponent research group has been active in the proposal of novel algorithms for pinpointing genetic modifications that can lead to enhanced production capabilities by using available genome-scale metabolic models [4;5], as well as in the development of computational tools that enable Biotechnology researchers to effectively use these methods [6]. However, a long road is still ahead in this field. In fact, available phenotype simulation algorithms are still limited to metabolic models, disregarding known regulatory information in the optimization processes and thus limiting the biological feasibility of the solutions. Furthermore, optimization targets only the task of identifying gene deletions towards the maximization of an objective function, while other interesting problems can be addressed. Available tools are also hampered by the lack of appropriate tools for the visualization and analysis of the provided solutions. In this scenario, the main aim of this project is to develop a set of computational tools to support ME, covering the main areas referred above: phenotype simulation using regulatory information, strain optimization and visualization/ analysis. These tools will be incrementally included in the ME framework being developed by the group that aims to become a reference computational tool for this community. Regarding simulation, the main efforts will be devoted to the development of new algorithms that are able to take advantage of integrated models that include information related to regulatory events together with the metabolic information. The approaches for strain optimization will not only extend existing algorithms, to handle objective functions provided by the new simulation methods, but will also address additional optimization tasks: the partial inhibition of gene expression (identification of the optimal level of expression for a set of genes) or the selection of the best set of new reactions to add to a target strain. To address these challenges, a repository of metabolic and regulatory models for several organisms and their components (e.g. genes, reactions, metabolites) will be created from a number of sources that will be suitably integrated. The development of multi-objective optimization algorithms will provide an additional contribution, since often the result of the optimization processes is more naturally viewed as a set of solutions, each representing a trade-off between potentially conflicting goals (e.g. maximizing biomass and the production of a given compound). The visualization of metabolic and regulatory models is an active research problem and it represents an urging need for the users of ME tools. Since a comprehensive solution to this task is out of the scope of this project, the aim here will be to seek the integration of our ME tools with relevant initiatives: Cell Designer[7] and Cytoscape[8]. In terms of analysis tools, we will look into studies that view metabolic and regulatory models as graphs, analyzing their topological properties, using measures from social network analysis [9] (e.g. node degrees, shortest paths, hubs and clustering). The idea is to study the differences in the topology between the wild type and successful mutant strains, leading to the discovery of interesting patterns. On the other hand, the concept of Elementary Flux Modes[10] is being used to study the space of functional capabilities of biochemical networks. The main limitation is the substantial amount of time needed to perform the analysis, but recent improvements[11] make it feasible for larger networks. In this project, the focus will rely on using EFMs to improve strain optimization algorithms and studying how these interact with regulatory information. Also, the development of tools to efficiently navigate, query and visualize these structures will be deployed.