With the development of life sciences, information on interactions between biomolecules has accumulated, and these interactions are showing such a complex aspect that they are called networks. It is thought that the dynamics that arise from complex network systems are the essence of biological activity. For networks that have been experimentally identified, clarifying the dynamics that arise from them is one of the most important issues in current biology. On the other hand, networks do not have sufficient information to determine their dynamics. They do not provide information of mathematical formulas or parameter values, that are necessary to determine the details of dynamics of systems. However, even if the dynamics themselves are not determined, there are important aspects of dynamical properties that can be extracted from the network. In this chapter, we will introduce a new mathematical theory that uses only information of regulatory networks to understand the behavior of biological systems. Namely, key factors for observing/controlling the whole dynamical system are determined from network structure alone. We also show an application of the theory to a real biological system, a gene regulatory network for cell-fate specification in ascidian. We demonstrate that the system was completely controllable by experimental manipulations of the key factors identified by the theory from the information of network alone.