Cellular phenotypes are intricate phenomena created by complex gene regulatory networks. Rather than remaining fixed in a given phenotypic state, cells move along a rough epigenetic landscape. Mapping back and forth between regulatory networks and epigenetic landscapes is a grand challenge for biological physics. I will present results from two studies of these networks in microbes. First, I will discuss a potential new source of epigenetic information in bacteria. Using a model that combines molecular thermodynamics with mass-action kinetics, we have shown that the build up of supercoiling due to transcription can introduce gene expression correlations within DNA topological domains. These correlations may give rise to a new layer of weak interactions in the bacterial regulatory network. Second, I will discuss recent results in understanding how the cell processes and stores information about gradients. Using a multiscale model of the yeast-mating pathway, we have shown how the bistable architecture of the mating pathway plays a crucial role in integrating directional information. We also hypothesize that the cell uses the cytoskeleton to actively store information about spatial direction.