The X-ray Crystollography Group in the Department of Biological Sciences at Purdue has pioneered a variety of methods, techniques and computer software for the determination of the structures of the macromolecules. It has achieved major advances in the study of virus structures and, over the years has developed a foundation of computer programs for film processing, phase determination, model building, structure refinement, and so on, leading to the solution and investigation of macromolecular structures. Computing is currently one of the bottlenecks in this research. The gap between the computing resources needed and the ones available increases as new radiation sources for X-rays become available. This group can maintain its position as the leader in the field by migrating their work to parallel supercomputers such as Touchstone Delta machine. The principal research issues we plan to address are: 1. Development and analysis of parallel algorithms for film processing, multiple small least squares problem, and huge (one million plus variables) sparse least squares problems. 2. Development of data structures suitable for macromolecular computations and which can be partitioned for parallel methods. 3. Development of geometric based methods for problem partition and mapping computations onto MIMD (Multiple Instruction Multiple Data) computers. This work will not only advance the understanding of parallel computing but it will also provide a practical tool to enable an order of magnitude increase in power for the computational biology. We expect this will lead to a qualitative improvement in the computational macromolecular biology.
