All the versions of this article:
October 12th, 2009
“Model-based investigation of microbial metabolism to interpret gene essentiality results, illustrated on Acinetobacter baylyi ADP1 metabolism.”
Maxime Durot (mail) supports his thesis on October the 12th in Évry. His research was carried out in the “Computational Systems Biology” team led by Vincent Schächter. A this time, Maxime has joined Claudine Medigue’s group.
Microbial metabolism has traditionally been investigated at two different scales: the finest involves characterizing individually each reaction occurring in the cell; the largest focuses on global cell physiology. While both scales have recently benefited from technological advances, combining them remains, however, especially complex as the global physiological behavior of a cell results from the coordinated action of a large network of reactions. Mathematical modeling approaches have yet shown recently that genome-scale metabolic models could help in linking both scales. In this thesis, we explore the use of such models to expand the knowledge of reactions with a specific type of high-level data: gene essentiality data, assessed using growth phenotypes of deletion mutants. We will use as model organism the bacterium Acinetobacter baylyi ADP1, for which a genome-wide collection of gene deletion mutants has recently been created. Following a presentation of the key steps and developments that have been required to reconstruct a global metabolic model of A. baylyi, we will show that confronting observed and predicted phenotypes highlight inconsistencies between the two scales. We will then show that a formal interpretation of these inconsistencies can guide model corrections and improvements to the knowledge of metabolism. We will illustrate this claim by presenting model corrections triggered by A. baylyi mutant phenotypes. Finally, we will introduce a method that automates the correction of inconsistencies caused by wrong associations between genes and reactions.
Key words : Metabolism ; Mathematical modelling ; Biological network ; Growth phenotype ; Gene essentiality