A biological cell forms an extremely complex system with a complicated metabolic reaction network. The functionality of the cell is controlled by the genome consisting of thousands of individual genes. Novel measurement technologies provide us with huge amounts of biological data in a quantitative form, but the suitable modeling methods to deal with the high dimensionality and large systems are still missing. In this article, a novel approach for complex system analysis, neocybernetics, is applied to modeling a biological cell. It is shown that under certain assumptions both the metabolic network and the gene expression network can be interpreted as a combined linear system. The analysis of this system from the neocybernetic starting points is then elaborated. As an example, the dynamics of yeast gene expression network are modeled according to the presented principles. 1

Abstract: When modeling truly complex systems, like ecologies, powerful modeling principles are needed. Neocybernetics studies distributed systems where there is no centralized control; self-regulation and self-organization emerges from local interactions and feedback structures. It turns out that the emergent patterns can be interpreted in terms of multivariate statistical methods: A cybernetic ecosystem carries out principal subspace analysis of the resource variations in the environment. The “forage profiles ” of individuals together span the directions of environmental variability. Because ecosystem following the neocybernetic adaptation can maximally exploit the resources in the environment, such strategy has evolutionary advantage. Such optimality properties motivate deeper analyses, too. 1.

The thesis at hand was written at the Control Engineering Laboratory of the Helsinki