|dc.description.abstract||The significant increase in the availability of postgenomic data has stimulated the growth of hypothesis-generating strategies to unravel the molecular basis of nature. The application of systems theory to biological problems emerged in the early 1970s, and yet the computational methods developed to model biological networks and analyse their functionality have been seldom used for understanding the neurogenetic basis of cognition. The main interests of this thesis are the application of computational models to microarray expression data for the identification and analysis of biological networks related with long-term potentiation (LTP), the cellular correlate of learning and memory in mammals. The models include the analysis of co-expression and studies of dynamical stability.
The thesis starts with the application of established methods on gene expression analysis on the available expression data from LTP in order to identify networks of closely correlated genes in their patterns of expression to ultimately pinpoint putative key regulators not identified previously by classical differential expression analysis. The thesis continues with the analysis of previously identified gene networks regulated 20 min, 5 h, and 24 h post-LTP induction. A dynamical stability analysis using weight matrices suggests that the early network has a significant sensitivity to perturbations compared with randomly generated networks of similar characteristics. In addition, using random Boolean networks, we study the differential sensitivity to perturbations of these networks and we find that our results are consistent with a model of LTP as a complex cellular switch. In such a scenario, earlier networks are dynamically more unstable than later regulatory networks, which are proposed to be responsible for the new homeostatic state reached by the stimulated neurons. Key genes responsible for the dynamic properties observed are identified and discussed. In particular, we found that Egr2, a member of the Egr family of transcription factors was crucial to the bistability observed in the early-response network. Other genes previously associated with LTP have a more modest contribution. A functional analysis of these networks is presented and integrated with previous knowledge on the molecular basis of LTP.||