Authors: D. Edwards, L. Yan, K. Olszewski and M. Donlan
Affilation: MSI, United States
Pages: 45 - 48
Keywords: high throughput genome analysis, comparative modeling, database annotation
The function of a protein is determined more directly by its three-dimensional structure than its sequence, it is therefore particularly useful to know the structure of a protein. This has been demonstrated by the many successes of structural biology. Initiatives are underway that use comparative protein structure modeling to generate structural data for all completely sequenced genomes through the use of an automated pipeline. Analysis of such vast amounts of structural data enables the development of a new type of 3D data that describes a given function. This data can then be used to predict the function of other novel protein targets and thereby provide a completely new approach to the identification of function. An alternative approach to using 3D information to identify function is through protein threading. An example of a protein threading algorithm which uses predicted secondary structure to identify the putative fold of a given protein sequence is given. Integration of this type of three-dimensional protein information with genomic data can lead to a much clearer understanding of a protein's function and further focus the mining of chemical databases for structure-based drug discovery. A description is given of the current state of homology-based data generation technologies. The developing techniques for characterization of active sites are discussed and evaluated in the light of the large volume of genomic data now available. How this data may be readily integrated into a drug discovery program is also highlighted.