De Shaws Supercomputer

PREV 2 of 2

Characteristically, Shaw has been mostly mum about Anton, referring the inquisitive to a technical paper on the project in the journal Communications of the Association for Computing Machinery.

His computer uses the massively parallel computing technology that Shaw helped develop at Columbia in the 1980s. Anton simultaneously runs 512 specialized processors called "application-specific integrated circuits."

Unlike other supercomputers that have more general-use applications, including weather forecasting, these processors are specifically designed to calculate the three-dimensional characteristics of molecules.

Shaw's team could use Anton to solve one of the most perplexing mysteries of molecular life: how proteins, the building blocks of life, each acquire a distinctive three-dimensional shape that allows them to perform millions of functions in a living organism.

Proteins, which include enzymes, hormones, and the collagen in bones and skin, are made in cells according to instructions from DNA. They're strands of amino acids bunched up like wads of string into distinctive shapes and held together by subtle physical forces that are still poorly understood.

Current supercomputers, including IBM's BlueGene/L and Stanford University's Folding@home (which uses legions of idle laptops to increase computing power), can take thousands of hours to simulate the folding of a single protein. Even then, these computers can create simulations of functions in molecules that last only a billionth or a millionth of a second. Scientists must then validate the findings.

Anton could run simulations going up to 1,000 times longer, allowing scientists to get much closer to what really happens when, say, a protein folds. "If you can do 1,000 times longer, real proteins come into play," Shaw reportedly said in a lecture at Stanford in 2006.

The more that scientists know about proteins and other critical molecules in the human body, the more precise they can be when developing drugs.

"He's making a big step forward with this," Benoit Roux, a biophysicist at the University of Chicago, told the New York Times.

Roger Brent, director of the Molecular Sciences Institute in Berkeley, California, suggested in the Times article that scientists may not know what such a powerful computer is capable of until they use it.

He pointed out that the original Anton—Van Leeuwenhoek, who perfected the microscope in Holland in the 17th century—didn't know that protozoa and other single-cell organisms existed in pond water until he trained his newfangled lenses on a sample.

Shaw also is a major investor in Schrödinger, a chemical and bio-physical simulation software business that could benefit from Anton's new technology.

PREV 2 of 2