Computer Science, Chemistry
Mathematics and Statistics
The basis of all biological material is proteins. Therefore, a better understanding of the actual confirmation of a given protein is key to understanding its biological function. Protein structures are represented in experimental data from either X-ray crystallography or nuclear magnetic resonance (NMR) spectroscopy. The data from NMR analysis, which was the focus of this study, gives a number of unique three-dimensional models, which represent the possible conformations of the protein, but does not take into account all of the dynamics of a constantly moving protein. The data collected from the analysis of each protein is stored in the Protein Data Bank (PDB) for easy access by anybody in a specially formatted text file. All the PDB files from X-ray crystallography consist of coordinates corresponding to the location of each atom in the protein as well as a B-factor value showing the variation of the position of each atom due to temperature changes. However, PDB files from NMR spectroscopy contain a series of models with coordinates only. The problem that this project addresses is representing all the models from the NMR data in a single model that show the dynamics of the protein where it is implied from the original models. A single structure with associated radii of variation for each atom is computed from data given in all models. This single structure is optimized to find the best fit to one of the original structures with the dynamics of the protein. Results from several proteins of varying structures are shown.