Courtesy of Folding@Home.

Proteins can fold in a myriad of different ways, which affects how they function inside living organisms. Vijay Pande from Stanford University, who led the Folding@Home project, likens protein folding to a ball of wool – determining the shape of the wool in the ball is a massive computational exercise. "What we needed was a single computer that was 100,000 times faster than what we had," he says, recalling the challenges facing his research team when they were broaching protein folding calculations in 1998. "Instead we got 100,000 computers."

Pande emphasizes that the two are not interchangeable, and multiplying the number of computers only helps if the calculation can be broken down into several parts. "For example, can 100,000 students take an hour exam, working together, and complete it in 0.036 seconds? That would be a challenging feat," he says.

While there were a few projects using volunteer computing back in 1998, devising a scheme to use this unusual computational resource was “the real challenge”. Having overcome this initial difficulty the approach is now being deployed to use idle smartphone processing time as well.

Number crunching to look for cures

When proteins fold wrongly disease can develop. Many of the protein-folding calculations by Pande’s team focus on understanding Alzheimer’s disease. "Alzheimer's is such a devastating disease, and it is occurring more and more frequently," says Pande. As well as the clinical significance of the problem, he also emphasizes the academic relevance of the topic, as Alzheimer’s poses important questions in biophysics.

The computational task may be daunting but it still offers advantages, as experiments in protein folding are notoriously difficult. "We hope to use novel computational methods to fill in the details that experiment cannot tell us," says Pande. However, a minimum level of accuracy is required for the calculations to be useful, and this adds to the computational burden.

"So far, we have developed lead compounds to fight Alzheimer's disease, made fundamental insights into the biophysics of protein dynamics, as well as insights into how key proteins behave in cancer," says Pande. With the rapidly increasing processing power of smartphones, harnessing idle time in these devices could provide a significant additional contribution to further advances in the team’s research.

The healing power of sleeping time

To attract and maintain volunteers for the project it is important to minimize any inconvenience. This can be quite straightforward for PCs, as they are unused for long periods at night, conveniently providing spare processing time for protein-folding calculations. The same was true of the PlayStation 3, when a release from Sony in 2007 allowed the protein-folding software to be installed on these devices as well.

With the limited battery life and ad hoc use of smartphones, identifying time that will not inconvenience the owner requires a little more thought. As a result the app will only make use of the phone processing power when the device is charging as well as idle. Again this is most likely to be while the owner sleeps.

As Pande points out, the computational methods, technology and data produced by the Folding@Home project have now been used by many other scientists throughout the world. "We're actively pushing in many areas," he says. "Our primary goal is to make a significant impact, with new drugs and new insights. We'll be making more announcements later in the year."

Full details of the project are available at