Specimens of chick cerebellum stained with osmium and reduced with potassium ferrocyanide. Image on right shows a higher magnification of a portion of image on left. The images were obtained in a study of the internal membrane structure of the Purkinje cell.
HPCC-supported advances in computer networks and visualization software are allowing scientists to control and share remote microscopes, telescopes, and other scientific instruments in an interactive, real-time fashion. These types of projects have opened the doors of the laboratory to a new concept Ð the "distributed laboratory" Ð which integrates laboratory equipment, high performance computing systems and data visualization tools over a high speed network, resulting in a more comprehensive and scientifically valuable investigative environment.
Researchers at the University of California-San Diego Microscopy and Imaging Resource have implemented a sophisticated, computer- controlled high-voltage transmission electron microscope (HVEM). Working in collaboration with staff scientists at the San Diego Supercomputer center and the Scripps Research Institute, the research group has coupled the HVEM via a high speed network to high performance computing systems and interactive visualization software running on scientists' workstations. The microscope is a unique resource, one of only a few such microscopes in the United States in use in biological science. More powerful than ordinary electron microscopes, it can accommodate much thicker laboratory specimens, yielding greater amounts of biological information. Using computer tomography and other visualization techniques, the images collected can be used to produce three-dimensional animations, allowing scientists to look at many previously uninterpreted areas of biomedical science relating biological function with structure.
The subjects of such study include the disruption of nerve cell components resulting from Alzheimer's disease, the structural relations of protein molecules involved in the release of calcium inside neurons, and the three-dimensional form of the Golgi apparatus, where sugars are added to proteins.
This project is an example of the application of diverse technologies to a single class of scientific problems. The capabilities of the microscope for scientific investigation will soon be extended, combining image data acquisition with computing resources to render, view, and animate the images for real-time analysis. By connecting the microscope to a high speed network, the instrument will someday be made available to investigators located in any geographic region, extending the accessibility of the resource to a broader scientific community in a collaborative environment.
Future plans call for extending this environment to the Apple Macintosh; implementing automatic focusing and calibration of the microscope (now handled by a human operator); developing remote image analysis tools; optimizing the tomography reconstruction code currently running on a Cray Y-MP system; and implementing the code on a parallel computing platform.
Scientists no longer need to be in the same room with their laboratory equipment. Instead, they can control it over the network from their desktop computers. Pictured is a scientist controlling a high-voltage electron microscope (useful for imaging thick three- dimensional biological tissues) from her Sun workstation.
The application of the operational concepts described above is not limited to microscopes. A research group affiliated with the National center for Supercomputing Applications (NCSA) in Champaign- Urbana, IL is investigating how to apply this model to real-time radio telescope observation. This group is seeking to demonstrate the feasibility of connecting the Berkeley-Illinois-Maryland Array (BIMA) radio telescope array, located at the Hat Creek Observatory in northern California, to NCSA supercomputers via high speed networks.
One of their needs calls for transferring a gigabyte-size observed visibility dataset from the telescope to NCSA for processing, and then returning the processed data to Berkeley for analysis on a workstation. In this way, the astronomer can judge the quality of the data, see if the signal is strong enough to proceed with the observations, judge whether the area of the sky being mapped is correct, and experiment with processing parameters. By connecting BIMA directly to BLANCA, one of five HPCC supported gigabit testbeds, and increasing the number of present antennae from three to six, the group expects to achieve a data transfer rate 10 to 100 times higher than current rates. This could allow real-time steering of the telescope, thereby greatly enhancing the value of an observational session and revolutionizing the way in which astronomy research is performed.