Section: New Results

Modeling through comparative genomics

Participants : David James Sherman [ correspondant ] , Rodrigo Assar-Cuevas, Nicolás Loira.

Using comparative genomics to inform mathematical models of cell function is a central challenge of the MAGNOME research program. Emmanuelle Beyne developed in silico methods for predicting protein complexes, one form of protein-protein interaction that provide the building blocks of cell machinery. These predictions were compared to experimental results from gel electrophoresis. This work was extended in a large-scale experimental study using quantitative proteomics and expression data, during a long-term visit to Prof. Steve Oliver's lab at Cambridge University. Florian Iragne has refined his methods for subtractive modeling of biochemical pathways, using his algorithmic framework for policy-directed graph extraction to identify cases of pathway loss through search for correlated gene losses. Nicolás Loira has used a large dataset of protein families from the Génolevures complete genomes and sub-partitioned it through clustering methods to obtain reliable indications of enzyme conservation in nine species. The resulting determination of enzyme conservation is mapped to biochemical reaction models and used to infer stoichiometric models that are currently being evaluated through comparison with experimental results produced by Prof. Nicaud's group at AgroParisTech.