Hybrid Human and Machine Computing
Starting with SETI@home in 1999, numerous distributed computing applications using spare cycles of PCs and other computing devices have been able to inexpensively tap massive computing resources. According to Wikipedia, SETI@home can tap 424 TeraFLOPS (Trillion FLoating point Operations Per Second), compared to a sustained rate of 478 TeraFLOPS for IBM's Blue Gene, one of the world's fastest supercomputer. Folding@home hit 1000 TeraFLOPS in 2007 by tapping into the power of PlayStation 3 clients and Graphics Processing Units of video cards.
Technology Review's Biologists Enlist Online Gamers discusses the Rosetta@home project that attempts to predict the three-dimensional structure of proteins. Although genome analysis has helped identify the molecular sequence of proteins, how the amino acid sequences in proteins are folded significantly affects protein function. Even simply proteins have many possible structures - finding "best one is regarded as one of the hardest problems in biology today".
Rosetta@home displays the 3D protein structure on a screensaver - some users reported that they could see opportunities to improve on the 'fit' of the structure. This led David Baker to work with Zoran Popović (a game designer at the University of Washington) to develop Foldit, which "attempts to predict the structure of a protein by taking advantage of humans' puzzle-solving intuitions and having people play competitively to fold the best proteins." It combines the ability of computers to sift through many alternative structures and visually represent form and function with the human ability to recognizing patterns.
As a game, the design is challenging since neither the goal nor the path is well understood, and in fact is developed by the users as they use Foldit. Analysis of player interaction with Foldit may reveal strategies that can then be incorporated into Rosetta@home, an interesting form of co-operative competition between silicon and biology.
