PROJECT LEADERS
Mark Ellisman, Fran Berman
UC San Diego

Carl Kesselman
University of Southern California

Rich Wolski
University of Tennessee

PROJECT MANAGER

Martin Hadida
UC San Diego

PARTICIPANTS

Jim Hayes, Salman Khetani, Stephan Lamont, Albert Lawrence, David Lee, Maryann Martone, Steve Peltier, Shava Smallen, Mona Wong
UC San Diego

Mei-Hui Su
University of Southern California

Gwen Jacobs, Sandy Pittendrigh
Montana State University

Amarnath Gupta, Tom Hutton, Bertram Ludäscher, Reagan Moore, Mary Thomas
SDSC

Chandrajit Bajaj, Arik Shamir
University of Texas, Austin

Martin Swany
University of Tennessee

www.npaci.edu/Alpha/Telescience

www-ncmir.ucsd.edu

Demonstrating the Synergy of the Latest Grid Technologies for Telescience

WHERE ARE THEY NOW?

WHERE ARE THEY GOING?

SHOW AND TELL

t began with telemicroscopy, the use of high-speed, continent-spanning data networks and modern computers to control a microscope in one city from a laboratory in another. Mark Ellisman, director of the National Center for Microscopy and Imaging Research (NCMIR) at UC San Diego, has pioneered the technology with software and hardware that make the 400,000-volt electron microscope at NCMIR available to researchers across the country. Since a specimen can go almost anywhere by overnight courier, telemicroscopy enables scientists to do productive research at first-rate facilities almost immediately instead of spending days in preparation and travel. Now NPACI's Telescience for Advanced Tomography Applications alpha project is extending researchers' capabilities even further, creating a system that not only will acquire images on a remote electron microscope, but also will use the most effective computing resources--wherever they may be--to derive 3-D structures of neurobiological specimens, abstract important information, and deposit the most useful representations of cell-level brain structures into a database to create a library for the scientific community, all accomplished via an easy-to-use Web interface.

"Studies of the tiny component structures of nerve cells can give scientists fundamental insights into how the brain and nervous system function," said Martin Hadida, manager of the Telescience project. "But the imaging instruments that can reveal these structures are few and expensive. Analyzing the data after images have been acquired presents further problems, since 3-D reconstructions from 2-D tomographic data require significant computing power. Ideally, scientists should have access to both capabilities at the same time for refining results and resolving structural ambiguities while they are using the microscope."

The Telescience alpha project is developing a coordinated, Web-based system for both telemicroscopy and data analysis of biological specimens. Data acquisition instruments will be linked to distributed parallel computational grid for analysis and refinement of the acquired image data. Under a framework now being developed for a multiple-scale brain-mapping project, acquired data sets will be inserted into databases that will grow to become libraries of cell structure information. The instrumentation control interface will be based on the ubiquitous Web; an advanced user interface environment will enable researchers to locate, render, display, and analyze surface-based 3-D renderings of cellular structures from the data libraries.

The project's most ambitious goal is intelligent, remote steering of imaging instruments. In this scenario, quantitative information from graphical models of refined data (based on results from remote analyses and comparisons with related entries in the database) from the microscope during an imaging session will lead the system to generate control commands to the microscope to more accurately re-acquire data from the specimen or to reorient the specimen to a new viewing angle to fill in missing data or to resolve ambiguous structures in the reconstructions.

WHERE ARE THEY NOW?

Ongoing collaborations between neuroscience researchers at UC San Diego's NCMIR and computer scientists at UC San Diego, the Information Sciences Institute (ISI) at the University of Southern California, and elsewhere, have been attacking the various components of teleinstrumentation, interactive analysis, and data analysis and archiving using distributed resources. "Integrating these collaborations eventually will create a seamless and transparent grid-based environment for scientific discovery," said Ellisman, the leader of the alpha project.

The effort involves a diverse collection of applications, technologies, and researchers: Globus, with Carl Kesselman of ISI; Application-Level Scheduling (AppLeS), with Fran Berman of UC San Diego and Rich Wolski of the University of Tennessee; Network Weather Service (NWS), also with Wolski; the SDSC Storage Resource Broker (SRB), with Reagan Moore of SDSC; Interactive Collaboration Environments, with Chandrajit Bajaj of the University of Texas, Austin; Telemicroscopy, with Ellisman and the NCMIR research team; and Multi-Scale Databases, with Ellisman and Gwen Jacobs of Montana State University.

Initial Telescience efforts modified the telemicroscopy code package to run on Globus, a program for distributed computing environments that is part of the Metasystems thrust area and one of the major elements of NPACI's grid-based computing efforts. Globus tomography software has been in production use by scientists at NCMIR on a daily basis throughout the past year. Several researchers from other laboratories in the U.S. and abroad have begun to use the telemicroscopy system on a regular basis.

More recent work has focused on grid computing to provide data analysis and information library capabilities. The project has integrated telemicroscopy, an early parallel tomography code, and the Globus, AppLeS, and NWS metasystem tools. Work continues in visualization, computational methods for large data sets, and distributed multi-scale databases, with integration of these components planned for the course of the next year.

WHERE ARE THEY GOING?

Future efforts on the Telescience project will extend its discovery capabilities, using products and resources of the Data-Intensive Computing Environments (DICE) thrust area. Specifically, the SDSC SRB will help federate databases populated with raw and refined data acquired by the telemicroscopy system. Cell-level data from tomography will be integrated with a multi-scale database of brain structures derived from other NPACI Neuroscience thrust projects, including the brain-mapping project at UCLA's Laboratory of Neuro Imaging (LONI) and neuron modeling efforts at Montana State University.

Eventually, as scientists use the NCMIR microscope and get feedback about their data volumes, they will also be able to query the federated databases for comparisons against similar structures and to visualize elements of the cell--or the cell in its entirety--for more thorough analysis. As envisioned, the whole process will occur quickly and transparently enough to inform and guide researchers' decision-making processes regarding specimen data acquisition. The cell-level data sets from automated tomography will be available to researchers who use computer-based modeling tools such as MCell, GENESIS, and NEURON. As part of the NPACI Neuroscience thrust area work, GENESIS and MCell are being expanded to operate in conjunction with the PACI computational infrastructure.

The Telescience alpha project is collaborating with NPACI's HotPage and the more generalized Grid Portal Toolkit (GridPort) to develop telescience portals--Web-based systems for computing, authentication, security, resource discovery, and data movement on the PACI grid. The GridPort software and architecture are designed by the Computational Science Portals team at SDSC, led by Mary Thomas. GridPort allows application developers to build Web-based portals to the grid while hiding the complexity of managing grid resources, allowing users to concentrate on the scientific components of the problem without worrying about the underlying portal infrastructure.

In an effort to accommodate technologies that will dominate computational grids in the near future, the Telescience alpha project has established a network at UC San Diego based on Internet Protocol version 6 (IPv6), and has ported most of its software to this next-generation protocol. Internet engineers expect IPv6 to gradually replace the current IPv4 protocol over the next several years.

SHOW AND TELL

Ellisman and Hadida conducted live telemicroscopy demonstrations at the INET 2000 conference in Yokohama, Japan, in July 2000 that successfully exhibited IPv6 capability. NPACI researchers also carried out telemicroscopy experiments this year with the Osaka University Research Center for Ultra-High Voltage Electron Microscopy, in which microscopes on both sides of the Pacific Ocean were operated by remote users. Similar work has begun with researchers at the Karolinska Institute in Sweden.

"We will be demonstrating a functioning system at SC2000 that is being used by researchers to accomplish scientific objectives in the real world," Ellisman said. "While we have what might be called production versions of these telemicroscopy systems in place, we are enhancing their capabilities through collaboration with network engineers and computer scientists. This is a project in which leaders in computer science are developing new technology with the immediate pull of a very demanding set of requirements from the application scientists." --MG *