 |
The impact of HPCC on all phases of biomedicine, ranging from basic research in molecular and cell biology to scientific and clinical visualization and modeling, has been impressive. The expanded capability of computer hardware and software means that scientific and clinical Grand Challenges not contemplated a few years ago can now be vigorously addressed. The examples summarized below represent only a small fraction of the recent accomplishments in research supported by various agencies. |
 |
Structure of the light harvesting complex II, a pigment-protein complex capturing sunlight in the photosynthetic membranes of the purple bacterium of Rs. molischianum. This is the first structure obtained through a combination of high performance computer modeling (to determine scattering phases) and x-ray diffraction (to determine scattering amplitudes). The approach promises to extend the applicability of x-ray scattering in structural biology and to simplify structure analysis. |
Several models for protein folding have been developed that may become sufficiently robust to handle an arbitrary protein -- an invaluable aid to the pharmaceutical industry in their quest for new and more disease specific drugs. A new method has been developed that combines computer modeling (1 2 3) with x-ray scattering to determine protein structure. This approach yields an accurate structure for a light-harvesting protein. Three protein design teams are collaborating on the use of networking and real-time high performance computing to design de-novo proteins. Researchers are also investigating how damage to human DNA is recognized by repair enzymes and how repairs are made. Researchers have developed a three-dimensional description of processes that are templates for structure-based design of drugs capable of regulating the intricate intracellular signaling system. This work points to ways to attack human ailments ranging from birth defects to spinal cord injuries.
Other important high performance computing applications in structural biology include algorithm development to model and manipulate DNA, search and analyze protein sequence data, and determine energy profiles of important biological molecules such as enzymes.
In one research center, high performance computer systems that store and analyze knowledge (1 2 3 4 5) about molecular biology (1 2 3 4 5), biochemistry, and genetics have been created. Specifically, parallelized search methods for sequence comparison have been developed that enable researchers to compare unknown sequences with every known sequence in under one minute, thereby facilitating the gene discovery process. This resource has over a half-million DNA sequences in its public database, which is accessed by over 20,000 different sites daily.
 |
A 23-degrees-of-freedom model for simulating human motion on earth and in space. The algorithm for this model achieves nearly linear scaling on a MIMD parallel architecture. |
 |
 |