Genomics, complexity and drug discovery: insights from Boolean network models of cellular regulation

The completion of the first draft of the human genome sequence has revived the old notion that there is no one-to-one mapping between genotype and phenotype. It is now becoming clear that to elucidate the fundamental principles that govern how genomic information translates into organismal complexity, we must overcome the current habit of ad hoc explanations and instead embrace novel, formal concepts that will involve computer modelling. Most modelling approaches aim at recreating a living system via computer simulation, by including as much details as possible. In contrast, the Boolean network model reviewed here represents an abstraction and a coarse-graining, such that it can serve as a simple, efficient tool for the extraction of the very basic design principles of molecular regulatory networks, without having to deal with all the biochemical details. We demonstrate here that such a discrete network model can help to examine how genome-wide molecular interactions generate the coherent, rule-like behaviour of a cell - the first level of integration in the multi-scale complexity of the living organism. Hereby the various cell fates, such as differentiation, proliferation and apoptosis, are treated as attractor states of the network. This modelling language allows us to integrate qualitative gene and protein interaction data to explain a series of hitherto non-intuitive cell behaviours. As the human genome project starts to reveal the limits of the current simplistic 'one gene - one function - one target' paradigm, the development of conceptual tools to increase our understanding of how the intricate interplay of genes gives rise to a global 'biological observable' will open a new perspective for post-genomic drug target discovery.