Abstract for the talk at 28.01.2014 (15:15 h)VW Seminar
Bruno Pace (Bioinformatik, Universität Leipzig)
Evolution of regulation and frequency landscapes
Boolean networks have been successfully applied to model gene regulatory networks. Motivated by biological systems' computing ability, we started to couple boolean with metabolic networks to observe their evolution. To do that, we numerically implemented a fixed size population of organisms which divide upon accumulating biomass. This is achieved depending on their correct switching of reactions and the consequent production of one or more target molecules, similar to a percolation in the metabolism of that organism. A biomass penalty proportional to the number of enzymes being produced is applied to them, in order to avoid the trivial solution (all reactions on). Each organism has its own boolean network and upon division it produces an exact copy and a mutant one. Some metabolites in the metabolic network are also present as ingoing nodes in the boolean network, acting thus as sensors of a varying environment. Conversely, some boolean nodes represent enzymes in the metabolism and are responsible for turning the corresponding reactions on. This effectively couples both networks of each individual. The topology of the metabolic network is shared by the whole population, and different environmental conditions are proposed as different adaptation tasks. This process can be described as a frequency landscape, where each genotype is associated to a frequency of division. Also, the frequencies (which play the role of fitnesses) change according to different environmental conditions. We will show that this population of oscillators may behave differently from expected when compared to a Fisher Wright process where fitness is identified with average number of offspring.