Systems biology attempts to reconcile large amounts of disparate data with existing knowledge to provide models of functioning biological systems. The cyanobacterium Cyanothece sp. ATCC 51142 is an excellent candidate for such systems biology studies because: (i) it displays tight functional regulation between photosynthesis and nitrogen fixation; (ii) it has robust cyclic patterns at the genetic, protein and metabolomic levels; and (iii) it has potential applications for bioenergy production and carbon sequestration. We have represented the transcriptomic data from Cyanothece 51142 under diurnal light/dark cycles as a high-level functional abstraction and describe development of a predictive in silico model of diurnal and circadian behavior in terms of regulatory and metabolic processes in this organism. We show that incorporating network topology into the model improves performance in terms of our ability to explain the behavior of the system under new conditions. The model presented robustly describes transcriptomic behavior of Cyanothece 51142 under different cyclic and non-cyclic growth conditions, and represents a significant advance in the understanding of gene regulation in this important organism.