Current approaches to the analysis of protein structure are time-consuming due to the lack of integration and incompatibility of data among the processing tools utilized. In addition, the enormous amount of experimental data generated is typically stored in flat file formats that cannot easily be managed or queried. Available databases that support the storage of protein structure data use data models that lack explicit definition of relationships between entities in the model necessary to support comprehensive querying and analysis. This paper presents a logical data model (LDM) for molecular structure determinations that utilize nuclear magnetic resonance (NMR) spectroscopy as the methodology. The LDM serves as a core component in an integrated framework for the comprehensive analysis of protein structure. Two ?directed? data mining queries about protein structure were used to validate the utility of the LDM in a database implementation.