Published online 14 October 2010 | Nature | doi:10.1038/news.2010.541


Supercomputer sets protein-folding record

Faster simulations follow protein movements for longer.

Anton, a special-purpose supercomputer, is capable of performing atomically detailed simulations of protein motions over periods 100 times longer than the longest such simulations previously reportedSimulating protein movements using Anton could aid drug design.SCIENCE/AAAS

A specially designed supercomputer named Anton has simulated changes in a protein's three-dimensional structure over a period of a millisecond — a time-scale more than a hundred-fold greater than the previous record.

Proteins are strings of amino acids that fold into intricate structures, which largely determine a protein's function. Understanding how and why proteins take on specific shapes has long been a goal of structural biologists, but previous computer simulations were too short to fully model the process.

Much of the effort has been directed at simply predicting the end product — a protein's final structure — on the basis of an amino-acid sequence. Anton goes further by providing a rare, detailed glimpse into the dynamic life of a protein as it folds and unfolds, twists and wriggles.

"We grew up with the view that a folded protein is static like a rock, but in fact it's not," says structural biologist David Eliezer of Weill Cornell Medical College in New York, who was not involved in the study. "It's highly mobile. It breathes and transitions between conformations."

Computing power

Anton was created by researchers at D.E. Shaw Research, an independent research institute in New York founded by David Shaw, formerly a professor at Columbia University in New York. Shaw abandoned academia in 1986 to work on Wall Street, eventually starting his own hedge fund. The fund was a success: in 2009, Shaw was number 123 on Forbes' list of the 400 richest Americans, with a net worth of US$2.5 billion.

Shaw returned to research in 2001, using his wealth to establish a research institute where he would be free to pursue his passions without relying on federal grants. He decided to tackle basic questions surrounding protein dynamics.

“We grew up with the view that a folded protein is static like a rock, but in fact it's not.”

He named Anton after Antonie van Leeuwenhoek, the seventeenth-century 'father of microbiology' who was the first to use microscopes to study microbes. From chips to algorithms, the computer is designed to one end: modelling particle–particle interactions.

Anton's simulations are based entirely on physical models of the forces among atoms in the protein and its surrounding water molecules. The computer divides time into tiny steps, each perhaps a femtosecond long, and determines how the atoms will move in each period by calculating the forces between all the pairs of atoms in the system.

Test run

To test Anton's ability to model protein dynamics, Shaw and his team selected two proteins that have been studied experimentally for a long time1. One is a protein fragment called a WW domain, and the other a small protein called basic pancreatic trypsin inhibitor.

The simulations revealed how the proteins changed as they folded, unfolded and folded again. "The agreement with experimental data is amazing," says Chandra Verma, a computational structural biologist at the Bioinformatics Institute of the Agency for Science, Technology and Research in Singapore.

“We could go longer, but at a certain point you lose patience.”

Simulating the basic pancreatic trypsin inhibitor over the course of a millisecond took Anton about 100 days — roughly as long as computers spent toiling over previous simulations that only spanned 10 microseconds. "We could go longer," says Shaw, "but at a certain point you lose patience."

Such simulations are highly dependent on the quality of the underlying physical models. Before Anton's success, there was much debate about how well those models would hold up over long time scales, says Verma. "The stability of the simulations over this new time scale has proven the sceptics wrong," he says.

Protein misfolding

Understanding those movements could be critical for understanding basic cellular processes, and for designing drugs that target particular protein conformations, says Eliezer. Proteins may become unfolded before they are degraded, for example, or when they are prepared for transport across a cellular membrane. And proteins that are not properly folded can cause diseases such as cystic fibrosis and Alzheimer's disease.


"If you had a better understanding of protein folding, you could understand protein misfolding," says David Baker, a computational structural biologist at the University of Washington in Seattle.

There are likely to be limits to Anton's powers, however. Shaw is the first to acknowledge that the system needs to be tested using more proteins before researchers will know how robust it is. And Baker notes that it is not uncommon for large proteins to take seconds to fold — a time scale that may yet be beyond Anton's grasp.

Nevertheless, Eliezer is optimistic that Anton will provide a way to shortcut the painstaking laboratory work needed to experimentally catch a glimpse of protein structure intermediates. "People have spent agonizing amounts of time collecting data to get a glimpse of how proteins fold," he says. "The ability to capture even just some of this in the computer is amazing." 

  • References

    1. Shaw, D.E. et al. Science 330, 341-346 (2010). | Article | ChemPort |


If you find something abusive or inappropriate or which does not otherwise comply with our Terms or Community Guidelines, please select the relevant 'Report this comment' link.

Comments on this thread are vetted after posting.

  • #60868

    I live in my house's garage and in the ceiling is a big hole which is where the attic is (they're in the middle of remodeling, hence the hole). Anyway, whenever I look up at the hole, I think I see a figure for a fraction of a second, then I blink and it's gone.

Commenting is now closed.