Abstract

The development of the technology to synthesize new genomes and to introduce them into hosts with inactivated wild-type chromosome opens the door to new horizons in synthetic biology. Here it is of outmost importance to harness the ability of using computational design to predict and optimize a synthetic genome before attempting its synthesis. The aim of this thesis is to help enable the engineering of synthetic genomes of one prokaryotic and two eukaryotic cells by using quantitative genome-scale models. Here, I develop a novel methodology to the automatic design of synthetic genomes which is based on an optimization that computationally mimics genome evo- lution. First, I address the design of the genomic transcriptional network of Escherichia coli with adaptation to varying environments. Applying reverse- engineering methods to the large amount of transcriptomic and signalling data available for the bacterium, I seek to understand the design principles determining the regulation of its transcriptome. I find that E. coli genome could be reengineered in such a way that it has a simpler transcriptional reg- ulatory structure while still maintaining the global physiological response to fluctuating environments. These genomes are more sensitive and show a more robust response to challenging environments. Second, I address how virus reprogram the cellular chassis of their host assuming that exist mech- anisms by which virus are able to overcome the defenses exposed by the host and modify its gene expression on its own benefit. I develop a novel genome-scale quantitative model of transcriptional regulation of Arabidop- sis thaliana for exploring the landscape of possible re-engineered genomes. I find a core set of host genes whose knockout or overexpression resulted in predicted transcriptional profiles that minimally deviate from the ob- served in plants infected. I perform this search for a set of eight viruses for which transcriptomic data are available and compared the results among them. Third, I extend the computational methodology for genome redesign to address the fine-tuning of the tomato fruit agronomic properties. I apply reverse engineering computational methods to transcriptomic, metabolomic and phenomic data obtained from a collection of tomato recombinant in- breed lines to formulate a kinetic and constrain-based model that efficiently describes the cellular metabolism from the expression of a minimal core of genes. Based on the predicted metabolic profiles, a close association with agronomic and organoleptic properties of the ripe fruit could be revealed with high statistical confidence. The model was used for exploring the land- scape of all possible local transcriptional changes with the aim of engineering tomato fruits with fine-tuned biotechnological properties. In sum, our results demonstrate that automated computational methods can efficiently explore the fitness landscape of re-engineered genomes with desired specifications.