Research suggests that network topology and interaction logic play a role in determining the states that a network supports. This concept is being applied to the study of cellular differentiation and reprogramming through data-driven inference of boolean networks from transcriptomes. The approach may have implications for predicting cellular behavior.