SCIENCE & ENGINEERING NEWS
San Diego, CALIF. — Bob Sullivan reports that the world’s fastest commercial supercomputer will soon be devoted exclusively to solving the world’s most powerful puzzle. A new IBM supercomputer will be made available to companies trying to ask questions of the recently decoded human genome.
Hailed as the scientific marvel of our lifetime, decoding the DNA that holds the recipes for human biology is merely a first step. The Human Genome Project, a full map of human DNA, has created perhaps the most powerful and intriguing database ever. But analyzing that data – really gleaning the secrets hidden inside our DNA – is a mathematical mountain.
For example, diseases like breast cancer are often not caused merely by one defective gene, but by a combination of defective genes. Since there are an estimated 35,000 identified genes, uncovering which combinations are at the root of an illness is an onerous task.
How onerous? It would take a traditional computer 447 years to solve the first equations that were constructed, according to Michael Mott, a spokesperson for NuTec Sciences, Inc. That’s a long time to wait for the health miracles promised by the decoding DNA.
To the rescueEnter IBM and NuTec Sciences, Inc., which specializes in algorithms devoted to speeding up computations related to the human genome project. Software designed by NuTec lets scientists ask questions of the genome data much faster – shrinking the wait time from 447 years to about one month.
NuTec will announce it is purchasing a new IBM supercomputer that would be the second fastest in the world. The system’s massive computational power will let pharmaceutical and biotechnology researchers study gene combinations behind complex diseases.
“Sifting through a mountain of genetic data to find a four- or five-gene combination that may be a factor in a particular disease is like looking for a needle in a haystack,” said Peter Morrissey, the president of NuTec Sciences’ Life Sciences Division. “This powerful tool will allow researchers to discover key relationships among subsets of genes. We needed an extraordinarily powerful computing environment.”
The new supercomputer will be capable of performing 7 1/2 trillion calculations per second. That’s 600 times faster than the IBM computer that defeated chess champion Gary Kasparov in 1997. The machine – actually a cluster of computers – will consist of 1,250 IBM servers supported by 2.5 terabytes of memory, 50 terabytes of online disk storage and a high-bandwidth networking infrastructure.
The only one that’s faster is ASCI White, a government-owned machine at Lawrence Livermore Laboratories.
The machine will be a big boost to research companies, since there is trial and error involved in asking questions of genome data, and one-month waits for each answer have made the research painstakingly slow.
Leasing time to other researchNuTec will be leasing time on the computer to large and small biotechnology and pharmaceutical firms looking to analyze various diseases and their causes.
“For this work, supercomputing cycles is what you need,” said IBM spokesperson John Buscemi. “This is a community of people who can access this machine.”
============================================================