The success of genomic sequencing projects in recent years has presented protein scientists with a formidable challenge in characterizing the vast number of gene products that have subsequently been identified. NMR has proven to be a valuable tool in the elucidation of various properties for many of these proteins, allowing versatile studies of structure, dynamics, and interactions in the solution state. But the characteristics needed for proteins amenable to this kind of study, such as folding capability, long-term stability, and high solubility, require robust and expeditious methods for the identification and optimization of target protein domains. Here we present a variety of computational and experimental methods developed for these purposes and show that great care must often be taken in the design of constructs intended for NMR-based investigations.