CHAPEL HILL, October 31, 2007—Integrated models for disaster planning and management, virtual environments for research and decision support, and computing and visualization to reveal the functions of proteins and gene mutations linked to cancer are among the presentations that will be featured in the Renaissance Computing Institute (RENCI) booth at SC07 in Reno.

The SC07 exhibition will be held at the Reno-Sparks Convention Center as part of the SC07 conference, beginning Monday night, Nov. 12, through Thursday, Nov. 15. The RENCI booth is #3215.

In its second appearance at an SC conference, RENCI will highlight its work over the last year with federal and North Carolina-based agencies to use advanced technologies to improve disaster planning and response. A presentation on the Institute’s FEMA-funded effort to create new high-resolution floodplain maps for coastal North Carolina will kick off the RENCI exhibit at 7 p.m. Monday, Nov. 12. The presentation will cover the background and approach to this work, which is still in progress and is expected to use more than 500,000 compute hours on RENCI’s Blue Gene/L supercomputer, Ocracoke. It will include visualizations on a high-resolution tiled display wall showing a hypothetical tropical storm slamming into North Carolina’s Cape Fear River.

RENCI also will demonstrate its collaborative workspace for emergency managers, called NC-FIRST (10:30 a.m. Tuesday, Nov. 13), and its use of the Regional Hydro-Ecologic Simulation System (RHESSys) to examine drought conditions in the southeastern U.S. (11:30 a.m. Wednesday, Nov. 14).

In addition, the RENCI will booth will highlight two projects designed to help researchers understand the complex interactions of proteins and genes and their relationship to human diseases. A demonstration with researchers from the National Center for Supercomputing Applications (NCSA) will present MotifNetwork, a user-friendly environment for facilitating analysis of protein domains, or motifs. These subsequences of amino acids contained within proteins define their activity and mode of regulation and offer another focus of analysis for studies of gene function, gene interaction and gene and organism evolution. Another presentation will highlight 3D visualizations created to help scientists study protein-protein interaction networks triggered by exposure to ultraviolet radiation. The goal of the work, which is funded by NCSA, is to better understand the changes in DNA caused by exposure to sunlight that trigger the development of melanoma, the most serious form of skin cancer.

“The presentations in the RENCI booth illustrate that this conference has evolved well beyond showcasing the latest powerful hardware,” said RENCI Director Dan Reed. “High performance computing, networking, visualization and collaboration technologies are now essential tools used by scientists and by government decision makers. Our field is affecting broad areas of the human experience, including how we diagnose and treat diseases and how we respond to natural disasters. The RENCI booth will highlight some of these broad impacts.”

Other features of the RENCI booth will include:

• An overview of RENCI research into novel uses of the 3D Internet.  In a demonstration conducted live from the RENCI Second Life Island, RENCI researchers will provide a tour of the island and the prototype virtual disaster command center it has developed there. The command center features real-time 3D maps, the ability to interact with avatars from a wide range of government agencies, and views of real-time data from sensors, satellites, the National Weather Service, and other sources. The demonstration, called Exploring the Emerging 3D Internet: Virtual Environments for Social Computing and Situational Awareness, takes place at 1 p.m. Tuesday, Nov. 13. Those who miss the demonstration will have the chance to explore RENCI’s Second Life island and view a virtual poster session using guest logins and avatars that will be available in one of the booth kiosks.

• A variety of quality of service issues related to large-scale grid computing will be examined in several presentations. At 4:30 p.m. Tuesday, Nov. 13, RENCI will demonstrate Teresa, a runtime qualitative performance analysis framework, and its use in providing performance validation and diagnosis for a workflow of the Linked Environments for Atmospheric Discovery (LEAD) program. At 10:30 a.m. Wednesday, Nov. 14, RENCI will demonstrate its fault tolerance and recovery system and its research to integrate the system with the Virtual Grid Application Development Software (VGrADS) stack.

• Booth visitors will have the chance to use a Gigapixel Image Viewer to manipulate super high-resolution images with a total pixel density of 1.2 gigapixels. Many of the images are 360-degree panoramas consisting of about 150 images digitally stitched together. With a wireless mouse, users will be able to pan, and zoom in and out on the images without any loss of resolution. RENCI is exploring use of this technology in fields as varied as medical diagnostics and disaster planning. The Gigapixel Viewer will be introduced at 8 p.m. Monday, Nov. 12. SC07 participants will be able to explore the gigapixel imagery on their own in one of the booth kiosks throughout the exhibition.

A scientific paper written by a research team from RENCI and the National Center for Supercomputing Applications (NCSA) was named the best application paper at the recent IEEE International Conference on Bioinformatics and Bioengineering (BIBE 2007).

The paper, MotifNetwork: A Grid-enabled Workflow for High-throughput Domain Analysis of Biological Sequences, was one of more than 600 submitted on scientific applications and the only one to receive an award. Its authors are RENCI Research Scientist Jeffrey L. Tilson and Gloria Rendon, Mao-Feng Ger and Eric Jakobsson of NCSA at the University of Illinois.

MotifNetwork is an effort to build an integrated suite of grid-enabled workflows for high throughput domain analysis of protein sequences. Traditionally, bioinformatics research involves analysis of genes and gene products—most often, proteins represented as sequences of amino acids. The proteins are analyzed by comparing their amino acid sequences, however, proteins also contain subsequences that define their activity and mode of regulation. These subsequences are called domains and motifs. Analysis of domains and motifs, rather than of protein sequences, often is a more productive way to study aspects of gene function, and gene and organism evolution. The analysis is computationally intensive because the possible combinations of protein subsequences grows exponentially as the motifs and domains interact. 

MotifNetwork is the first user-friendly environment to facilitate motif/domain analysis. The workflow orchestration and enactment is handled by Taverna, software tools used to make workflows and distributed computing technology easy to use. MotifNetwork uses the Generic Service Toolkit (GST) to allow the workflow to use grid services and perform grid-based computations. Its end products are data, organized as matrices, and visualizations suitable for quick analysis.

The NSB has recommended that NSF select the University of Illinois at Urbana-Champaign and the National Center for Supercomputing Applications (NCSA), in collaboration with IBM, the Great Lakes Consortium for Petascale Computation, and partners around the country to design, deploy, and operate this petascale system, pending final approval of the decision by NSF, which is expected in late summer or early fall. RENCI is a member of the Great Lakes Consortium for Petascale Computation.  For more information, please see the NSF press release.