CDC using cluster to study smallpox

CDC smallpox Web site

The Centers for Disease Control and Prevention have enabled scientists studying the genetics of smallpox to crunch numbers faster than ever without spending hundreds of thousands of dollars on specialized supercomputer technology.

The Linux Networx Inc. Evolocity cluster links 20 nodes, each with two processors, allowing scientists to whip through bioinformatics problems such as smallpox genomic functions, which aid in new vaccine development, company spokeswoman Andrea Bingham said.

"It's imperative to get information crunched faster so [researchers] can get on with creating cures," she said.

Scientists sometimes use supercomputers to handle such processing, but the cluster, based on standard computer hardware and Linux software, was much less expensive.

Dale Nordenberg, associate director for informatics at CDC's National Center for Infectious Diseases, estimated that a cluster machine cost CDC $40,000 to $60,000 while a supercomputer would have cost $200,000 to $500,000.

CDC didn't consider buying a supercomputer because the agency could get the necessary capabilities from a cluster, he said. In addition to the savings, using a cluster frees up server space for other applications.

The agency's Biotechnology Core Facility Branch has been using the cluster since September 2002, and genomic functions that may have taken one computer two weeks to complete are finished in a day, Bingham said. The cluster is running programs such as the Basic Local Alignment Search Tool, or BLAST, which determines the functions of newly discovered genes.

"A need was identified for parallel processing," Nordenberg said. "What would previously take us overnight to do now takes us five minutes."

Bioterrorism threats have caused officials to re-examine the need for the smallpox vaccine, which hadn't been administered since the 1970s. Certain military officials and first responders have been ordered to receive the vaccine, which can have deadly side effects. Such concerns have led CDC officials to study new vaccines.

Also, by using the tools to cut, sequence and process slices of smallpox DNA data, CDC officials can identify variations in the disease. Nordenberg said officials then catalog the genome's hundreds of variations by collecting samples from the few known storage locations. Officials will then be able to identify where a certain strain came from if there is an outbreak of the disease.

"If we identify an outbreak of smallpox, we can identify the strain," Nordenberg said. "Then, you start to look for the link to trace it to the source."

The Linux cluster includes management tools that enable administrators to keep a close watch on each node. For example, a tool can monitor the computer's temperature and available memory, alerting administrators of any hazards and turning off the affected node, Bingham said.

"They can just let the tools run, and they can concentrate on the genome sequencing and the study and not what's happening on Node 12," she said.

Michael Swenson, life sciences analyst for IDC, said those types of tools were available three years ago, but the current tools allow the cluster to function more as a single system.

If the need for smallpox research changes, the cluster can be used for other research areas, said Eric Pitcher, Linux Networx's vice president of product marketing, government, industrial and life sciences. Similar genomic functions are performed on most diseases to identify strains and help develop vaccines.

"It's a good infrastructure for the bioinformatics," Nordenberg said.

Swenson said CDC was doing what most biotechnology companies are doing, and the use of cluster machines has increased dramatically in the past couple of years.

"These programs they are running are considered embarrassingly parallel," Swenson said. "These applications break up very nicely into parallel chunks, and you can get it done very much faster."

***

Quick results

The Centers for Disease Control and Prevention are using Linux Networx Inc. cluster computers to study smallpox genomics, reducing the time needed to run complex applications such as Basic Local Alignment Search Tool (BLAST).

BLAST determines the function of newly discovered smallpox genes by providing rapid search of DNA and protein databases. It also detects alignments, offering clues to the functions of the proteins.