Project : symbiose

Section: Software

Metabolic and genetic networks

Participants : Michel Le Borgne [correspondant], Anne Siegel, Anurag Adarsh.

For supporting our research in modeling and analysis of dynamical systems of biological metabolic and genetic interactions, a software has been developed and is still under development. This software, named Garmen (Graphical Analyzer for Regulatory and MEtabolic Networks), implements a first version of a computer model of interaction and the derivation of a graph model.

Models are specified using a declarative language. The user has to enumerate the various interactions of interest occurring in a cellular localization (cytoplasm, mitochondria, nucleus, etc). Localizations are implemented as scopes in traditional programming languages. A compiler builds the object based model and a graph generator builds a graph representation of the biological networks. Various tools allow for a flexible display of subnetworks of interactions. Types of biological interaction include signal pathways and gene expressions. The explanatory part of the software has been improved and allows one to extract from DNA-chips data the subset that do not fit with the model.

The generation of simulation programs for Matlab is a new feature of Garmen. Informations are added to the description of the biochemical network from which GARMEN generates a Matlab program implementing the rate laws of the reactions and the stoechiometric matrix [41]. This is the first step in the realization of a software platform allowing for the study of a biological network from various point of view. Following the same idea, the generation of SIGNAL code (a language developed at IRISA) in order to model the combinatorics of gene/signal networks is under study.

Technical improvements have also been made on GARMEN in order to increase its portability, an important issue for a future distribution. This software is available on demand to the correspondant.