IBM has announced the first commercial life science customer for its Blue Gene/L computer, the company’s next generation supercomputer.
The Computational Biology Research Center (CBRC) of Japan’s National Institute of Advanced Industrial Science and Technology (AIST) will use a Blue Gene/L system to conduct 3D protein folding simulations aimed at accelerating research into new drugs.
“One of our biggest research challenges is to apply data obtained from genome decoding to protein engineering and drug design” says Yutaka Akiyama, CBRC’s director. “The scale of simulation this requires cannot be done without the help of supercomputers.”
To that end, AIST will install a four-rack Blue Gene/L system next February. The system will offer a peak processing power of 22.8 teraFLOPS (22.8 trillions floating point operations per second). This is about 24 times more powerful than the CBRC’s current computer systems. IBM notes that the system will take up a fraction of the floor space of the current CBRC systems. To put the four-rack Blue Gene/L system’s processing power into perspective, if such a system were in use today, it would be the world’s third most powerful supercomputer, according to the most recent list (published in June) of the world’s top 500 supercomputers.
The specific AIST work that will be done on the new Blue Gene/L is studying the folding of proteins in fluids. “Seeing the dynamics (of a protein folding) has been unachievable,” says William Pulleyblank, director of exploratory server systems, IBM Research. Today, the best that can be done is to go into the lab and look at a protein’s final structure.
“At a microscopic level, you can never see the [folding] process while it is occurring,” says Pulleyblank. With the simulations run on Blue Gene/L, AIST will be able to better understand the dynamics that take place during the folding itself. “Think of a microscope letting you see what the eye can’t see,” says Pulleyblank. The simulations “let you see what the microscope cannot see.”
As part of the deal, IBM Research and AIST are exploring the possibility of doing some joint research using software IBM has developed to run protein simulations on Blue Gene/L.
Such research collaboration is in line with the overall Blue Gene project, where IBM considers early Blue Gene adopters more like partners than straight customers.
To date, there are three other publicly announced efforts where a Blue Gene/L system will be installed within an organisation in 2005: Argonne National Laboratory, the Dutch astronomical organisation ASTRON, and the Lawrence Livermore National Laboratory. And last month, the Mayo Clinic entered into a partnership with IBM to apply Blue Gene’s processing power to improve patient care and advance medical research.
One major IT distinction between the AIST and Mayo applications is that the AIST group will have a Blue Gene/L installed at its facilities, while the Mayo Clinic effort will use an on-demand computing service where IBM will own and maintain the physical Blue Gene/L system.