From Genome to Life:

QUENTIN Yves

LCB-IBSM CNRS, 31 Chemin Joseph Aiguier, Marseille 13402, France

title: Exploration of genetic context around an anchor gene using constraints framework

With the availability of complete genomes there is a real need to develop new algorithms to complement more traditional similarity-based methods for predicting protein function. Among them, methods that intend to predict functional links between genes or their products from the conservation of genomic context appear to be very promising. In previous works, this question has been addressed throughout the analysis of genes belonging to the same putative transcriptional unit. However, there is some evidence of conservation of functional related genes in the same neighborhood without this constraint. Therefore, we have decided to develop a method of finding such conserved gene clusters around an anchor gene from a large number of genomes but without restricting the analysis to genes in the same orientation and in a close proximity. We have used the framework of the temporal constraint satisfaction problem, belonging to the artificial intelligence field, to represent and resolve the problem. Our algorithm is very fast and can treat a large set of data. It was applied on anchor genes coding for proteins involved in the assembly of ABC transporter systems.