The Resource for Concurrent Biological Computing pursues research projects in Molecular Biology, Neurobiology, Magnetic Resonance Imaging (MRI) and other areas which are extremely demanding computationally or not even possible on today's computers. For this purpose the Resource explores the use of massively parallel computers, developing expertise and programs for the computational biologist. The Resource seeks access to the highest performance machines at National Centers. It will operate a Gflop machine for program development, testing and limited production runs as well as operate a scalable parallel computer linked to a virtual molecular environment. The Resource will also pioneer the use of networked workstations as single machines for large scale biological computing. The Resource will develop numerical algorithms for parallel machines, e.g., for efficient description of electrostatic forces. It will provide molecular dynamics programs running across networks and translate physical concepts into computer programs, e.g., for long-time simulation and protein folding. It will also develop a neural simulator for parallel computers. Resource staff carries out projects in modelling large biomolecular systems like water-membrane-protein systems, redox proteins, and protein-DNA interactions, and studies biopolymer electrostatics. They also simulate the development of brain maps in the striate cortex and the development of motor control, model cognitive processing through coherent neural firing such as image segmentation, and study MRI microscopy. Collaborative projects include new approaches to structure determination from protein sequences, DNA-protein interaction, photosynthesis, membrane systems, molecular recognition by the immune system, sequence data bases, computational MRI, and adaptive filtering in vertebrate sensory systems. The Resource will closely cooperate with the National Center for Supercomputing Applications sharing the major hardware and utilizing Center staff for machine operation and user services. The visitor and workshop program of the Resource will be administered by the Beckman Institute at the University of Illinois.
Showing the most recent 10 out of 371 publications
