The following is taken from WCG description of their project hope it helps.

The name of the computer program is Rosetta. It computes a "Rosetta score" that tells how properly folded a protein is as the program tries different foldings. To compute this score, the program considers the packing of amino acids within the protein according to many scoring rules. The lower (more toward the negative) the scores are, the better the folding.

The Rosetta score for the best folding that your computer has identified for a particular protein is shown as the "Min" value under the current Rosetta Score in the left half of your agent's application window. You can see snapshot pictures of the partially folded protein that your computer is working on in the right half of the window. The left side shows two other numbers that tell how properly folded the protein is thus far. The Environment score shows how well the central core of the protein has been packed together. The Pair score tells how well certain amino acids have been paired up with the correct counterparts. If a trial fold gets a worse score, then the Rosetta program tries to refold the protein a different way to see if it produces a better score. This is done millions of times for each protein. Scientists will look at the best scoring protein structures and use those in the next steps of their research.

> What is the difference between this project and other protein folding projects?

There are a number of distributed computing projects like this one, with the goal of predicting protein structures given an amino acid sequence. In fact, there is a very large effort that is using our software, Rosetta, as the core application (the Human Proteome Folding Project). All of the projects will benefit our understanding of protein folding and application to biology and medicine. Our project is similar to Predictor@home, in that we are trying to improve our methods by conducting research that is only possible with the computing power that a grid computing project can provide.

------------------

Folding@home studies the process of folding using molecular dynamics. Distributed computing is allowing them to do simulations at time scales (milliseconds) that are not possible without thousands of computers. We, on the other hand, are doing protein fold predictions and design, including protein complexes. The WCG project is applying our Rosetta software to the human genome and other genomes, in collaboration with our group. We are trying to improve the software (focusing on research, development, and testing).
For example, what we learn through our project may help the WCG in their next phase for high-resolution structure prediction.

Written bye David Baker on the message boards