A growing number of diseases have been linked to a surprising common origin: protein misfolding. Proteins are the agents, messengers and building blocks of the body. It was long thought proteins that had to have the correct chemical composition and be produced at the right place. It turns out that their shape also is of supreme importance. Even if a protein has the right composition, it still can be assembled incorrectly and folded the wrong way. This misfold can render the protein useless or even harmful for the body's biochemistry. Protein misfolds are the cause of many illnesses, such as Parkinson's disease, Alzheimer's, Mad Cow disease, CJD and many others.
Unfortunately, the way proteins misfold is not fully understood yet. Only by simulating millions of foldings with varying parameters can we hope to understand it. Computer models are available for simulating protein folding, but they are terribly CPU-intensive. Simulating the folding of a single, simple protein can take more than a day on a fast single-processor machine. This is the kind of problem that begs for distributed computing.
That's why researchers in Stanford created the Folding@Home (FAH) project, which is run as a non-profit project by Stanford University's Chemistry Department. About 200,000 volunteers all over the world already downloaded the FAH client (available for Linux, MacOS X and Windows). The client gets a unit of work from Stanford's servers, number-crunches for a while, uploads the results, rinses and repeats. Linux currently accounts for only 5 to 10 percent of the clients.
The project already boasts impressive results. Because the simulation results are publicly available, your machine's computations can benefit all interested scientists. If you don't have a fortune to donate or a Ph.D. in biochemistry, the best way to fight horrible brain-degenerating diseases such as Alzheimer's is to contribute to this project.