Résumé
Gene network inference is the task of reconstructing regulatory networks among genes from high-throughput (in particular transcriptomic) data. Here we introduce the main concepts of this rich and rapidly evolving field. In order to illustrate the basic principles of gene network inference we simulate gene expression patterns using two distinct computational models, Boolean dynamics and ordinary differential equations. These synthetic data are then analyzed with basic gene network inference methods based on information theory and correlation analysis. We emphasize that a careful distinction between the underlying network architecture and the effective network inferred from the dynamical patterns , similar to the interplay of structural and functional connectivity often discussed in Computational Neuroscience, may open up a new perspective on the data.