CHAPTER 38
PROBABILISTIC APPROACHES FOR INVESTIGATING BIOLOGICAL NETWORKS
Last decade saw a significant increase of high throughput experiments. As a major achievement, these novel techniques replicate the molecular experiments, which opens perspectives of quantitative behavior investigations. For illustration, it is now possible to define the concentration for which a protein (i.e., transcription factor) may activate a given gene. This information used to be considered as a limitative factor for producing accurate dynamical models of large biological regulatory networks [5]. Today, one must take it into account for building large quantitative models. Furthermore, high throughput experiments describe, as well, macromolecular processes via their temporal properties. Thus, biological processes can be summarized by the evolutions of their biological compounds over time (i.e., a succession of biological qualitative states or temporal patterns). Such experiments show temporal parameters that refine, in a natural manner, the qualitative models describing biological systems. However, these refinements, which present great biological interests, raise similarly several computational concerns. One is dealing with the complexity that originates from the large amount of experimental data. The challenge hence consists in trimming the experimental information at disposal for extracting the major driving compounds and their respective interactions within ...
