News blog

Video gamers take on protein modellers

foldit240.jpg

Gamers are starting to give biochemists a run for their money. An online computer game that challenges players to contort virtual proteins into shapes that resemble a protein’s genuine structure fared well in a competition to predict protein structures.

Foldit is the brainchild of David Baker and his team at the University of Washington in Seattle. In the game, players twist strings of amino acids to find energy-efficient three-dimensional shapes for those peptides. In nature, protein shapes tend to occupy low energy states, and Baker’s team hopes that the solutions the Foldit gamers come up with can help scientists determine shapes of proteins.

In a 2010 proof-of-principle paper published in Nature, Baker’s team showed that Foldit players do a good job finding more efficient shapes for proteins whose real shape had already been determined. Now, in a paper published in Nature Structural & Molecular Biology, Baker’s team reports how Foldit players performed in a biannual competition called CASP (short for Critical Assessment of Techniques for Protein Structure Prediction).


In CASP, scientists are challenged to predict the shape of proteins for which the real structure hasn’t been determined experimentally, but soon will be. Every two years, scientists meet to learn how their predictions compare to one another’s and to the real protein shapes. The 17-year-old competition has helped identify common stumbling blocks in protein-shape prediction and focus research efforts, its founders have said.

Baker’s team let the Foldit players loose on some of the shorter proteins from last year’s CASP9 competition. Their team was called the Void Crushers. One of their entries, for a protein made by the bacterium Pseudomonas aeruginosa, was among the most accurate predictions in the competition. In another challenge that involved coming up with a more efficient shape of a known bacterial protein, the Void Crushers performed better than any other entry.

Foldit’s biggest success so far came after CASP9, on an enzyme produced by a retrovirus called Mason-Pfizer monkey virus (M-PMV). A player who goes by the name ‘mimi’ came up with a shape that would be accurate enough to serve as the basis for determining the real shape of the protein based on X-ray diffraction measurements.

“The M-PMV structure had stumped scientists for a very long time before Foldit players made their breakthrough. This is the first example I know of game players solving a long-standing scientific problem,” Baker wrote in an email.

The structure could help the design of new anti-retroviral drugs, Baker’s team says. That’s more than a World of Warcraft player can say at the end of a Mountain-Dew-fueled, all-night gaming session.

For more on Foldit, check out this Nature video:

Image of Foldit screenshot from Cooper, S. et al. Nature 466, 756–760 (2010)

Comments

  1. Report this comment

    Tim Longhurst said:

    Love it! And if this story has put you in the mood for stories on gaming and health, you can’t go past this oldie but goodie… Surgeons may err less by playing video games: http://goo.gl/Vh7Bg

  2. Report this comment

    Keynutrition said:

    Great article as I am interesting in the science behind predicting protein structures, love the line about world of warcraft too, lol!

Comments are closed.