A biophysical study of evolution

Abstract

Biological organisms represent physical systems that usually respond well to perturbations and can maintain their functionality within the presence of noise and fluctuations, both in the environment and during the development. The Kaneko et al papers [23] contain theoretical and computational approaches towards understanding the evolution of robustness in noisy biological systems. In this paper, we implement Kaneko's model for evolution in a population of individuals containing gene regulatory networks and we analyze the way evolution changes the internal properties of the population. The population increases its fitness by selecting for individuals that can produce positive expression levels of a target subset of genes at the end of each generation, and the shape of the expression curve depends on the initial properties of the regulatory network. We observe that the expression levels of the rest of the genes stabilizes at either maximum or minimum values, with few remaining around 0; this represents their roles as either activator, inhibitor or neutral within the network. The population is subjected to strong selection in the first couple of generations, which results in the progenies of one individual sweeping over the population; mutations play a significant role in later generations. Due to this phenomenon, the genetic variance of the population always decreases; on the other hand, the phenotypic variance shows a peak before the 10th generation, after which it decreases close to 0. This indicates that the robustness of the system changes during evolution, and that different distributions in the fitness of the individuals translate into selection for different levels of individual robustness.