Evolving boolean networks to find intervention points in dengue pathogenesis

We use probabilistic boolean networks to simulate the pathogenesis of Dengue Hemorraghic Fever (DHF). Based on Chaturvedi's work, the strength of cytokine influences are modeled stochastically as inducement probabilities. Two basins of attractors are observed with synchronous updating; the Null Infection cycle attractor shows an expected cross-regulation of Th1 and Th2 cytokines corresponding to the homeostasis of an uninfected person, while the DHF Infection attractor shows the onset of DHF. With asynchronous updating, our model remains valid with clinical comparisons against qualitative changes in signal durations. In order to find intervention points that could prevent DHF, we design a genetic algorithm to shift the DHF attractor to the DF attractor basin by using the DF final state as the fitness measure. Our simulation results identify TGF-, IL-8 and IL-13 as the intervention points which are consistent with known clinical results to prevent DHF from occurring. Categories an...

• Type: Conference
• Year: 2006
• Where: GECCO
• Authors: Philip Tan, Joc Cing Tay
• Keywords: DHF Attractor, DHF Infection Attractor, GECCO 2006, Infection Cycle Attractor, Optimization