My laboratory pursues research in four major focus areas, protein folding, protein sequence analysis, molecular assembly databases and protein engineering. We undertake technology development in integrated databases, visualization tools, clusters and distributed computing to tackle leading problems in proteomics and genomics. In order to do all this, we have created a new integrated database system called SeqHound. SeqHound will be described at the API level and examples of the kinds of computations and results it supports, shown from our research efforts. SeqHound is the "back-end" database supporting the Biomolecular Interaction Network Database - BIND. BIND is the fastest growing database in Bioinformatics, and is currently being filled with 4 new types of data: high-throughput genetic interactions, profile-based interactions, high-throughput mass-spectrometry derived topology-free complexes, and interactions in atomic detail from three-dimensional structures. Using our bioinformatics infrastructure platform, we have been able to quantitatively compare both sets of mass spectromety results from large-scale yeast protein interactions published in the Jan 12 2002 issue of Nature by Ho et al., and Gavin et al. The combined dataset of all known yeast interactions together with a crude complex-finding algorithm gives a compelling glimpse of the structure of the eucaryotic nucleolus from Saccharomyces cerevisiae.