The San Diego Supercomputer Center (SDSC) at the University of California, San Diego, has in the last three years undergone a major reboot, remaking itself into a center of expertise on all aspects of "big data" research including genomics, one of the fastest growing areas of scientific study.
"We have in recent years become a lot more than a supercomputer center," SDSC Director Michael Norman told attendees earlier this month at the third annual X-Gen Congress & Expo, a four-day event focused on exploring the potential of established and emerging genomic technologies. "Our real expertise is now in all aspects of 'big data,' which includes data integration, performance modeling, data mining, software development, workflow automation, and more. We believe that data-enabled science is the beginning of a new scientific era."
SDSC's creation of a fully integrated "big data" environment already has led to several projects in the study of genes, and more are underway. "Our focus in genomic medicine is growing," Norman says.
"Next-generation sequencing of DNA and RNA are profoundly transforming biology and medicine, providing insight into our origins and diseases," according to Wayne Pfeiffer, a distinguished scientist at SDSC. "However, obtaining that insight from the sequencer data deluge requires complex software and increasingly powerful computers."
SDSC has an expanding repertoire of "big data"systems, the latest being Gordon, a unique flash memory-based supercomputer that is capable of storing 100,000 entire human genomes, while operating hundreds of times faster than conventional computers to study genetic data.
Genetic data creates many additional requirements regarding sharing and computing. The iDASH center (integrating Data for Analysis, Anonymization, and Sharing), under the leadership of Lucila Ohno-Machado, is the most recent National Center for Biomedical Computing funded under the U.S. National Institutes of Health Roadmap for Bioinformatics and Computational Biology.
Conceived as a collaborative computational environment to improve access to health data and software, iDASH provides biomedical and behavior scientists with access to a sophisticated, secure privacy-preserving infrastructure to contribute, integrate, and analyze their data, as well as potentially reuse data from others (given permissions set up by data contributors) and leverage other research results.
"The iDASH center addresses fundamental challenges to research progress by providing a secure, privacy-preserving computational environment in which researchers can analyze molecular, clinical, and behavioral data," Ohno-Machado says.
SDSC also has multiple collaborations with the Scripps Translational Science Institute (STSI), which has a dedicated 1-gigabit-per-second network connection to the center, along with 140 terabytes of online project storage. STSI has purchased time on SDSC's Triton Resource to conduct research on a number of projects.
One such collaboration is called the Human Tumor Study, which is using SDSC's Triton Resource to search for genome variants between blood and tumor tissue. Software used in this project includes the Genome Analysis Toolkit, the SOAPdenovo assembler, and various aligners such as ATAC, BLAT, and BWA.
Another collaboration involving SDSC, STSI, and others is called W115. In this project, Pfeiffer is using the Velvet and ABySS assemblers and the ATAC and BFAST aligners on the Triton Resource to study the full genome sequence of a 115 year-old woman to determine how many mutations occur in a long, healthy lifetime.
Further collaborations between SDSC and other genomic institutions including STSI are expected, Norman says, noting that Gordon and its data storage facilities have the bandwidth needed for such research. "The end goal here is to develop a rapid learning system for guiding individual therapies, and SDSC is now set up assist in reaching that goal," he says.
In addition to Gordon, which went into production earlier this year, SDSC operates Trestles, designed to enable modest-scale and gateway researchers to be as computationally productive as possible, and the Triton Resource, a medium-sized data-intensive compute cluster primarily for UC San Diego and UC researchers.
All three computer systems, for example, are integrated into four tiers of specialized data storage, which is crucial for genomics and other researchers who need to sift through massive amounts of data. SDSC's data storage facilities include:
SDSC's four tiers of specialized storage are all interconnected for use as needed. "One can access any storage from any of these systems and build workflows that hop from one system to another," Norman says.
SDSC's Gordon and Trestles systems and their storage systems are available for use to any researcher or educator at a U.S.-based institution and not-for-profit research through the U.S. National Science Foundation's Extreme Science and Engineering Discovery Environment (XSEDE) program. Industry-based research time and storage is also available. Industry researchers interested in using SDSC's resources or expertise should contact SDSC's Ron Hawkins.
The X-Gen Congress & Expo was held March 5-8 in San Diego.
No entries found