Atomic scale interactions, although short-ranged, none-the-less, result in emergent phenomena that scale several orders of magnitude in both length and time. The development of tractable computational means of studying self-assembly, while retaining atomic-scale physics is a challenging problem that has seen decades of research efforts directed at it. In this dissertation, I apply a recently developed coarse-graining framework based on an informatics quantity known as the relative entropy to develop coarse-grained models of peptides. I use these models to study the thermostability and self-assembly behavior of peptides near an interface, and to develop rational design principles for modulating the structures they form. Further, I implement a larger-scale coarse-graining technique to elucidate the underlying cause of emergent chiral lattice structures in systems of quasi-2-dimensional colloidal particles, and put forth a general framework for designing colloidal particles that form chiral self-assembled phases. This dissertation applies powerful coarse-graining strategies to elucidate bottom-up design principles that have tremendous applicability to the development of biotechnologies and advanced materials.