The science behind protein folding
Proteins are tiny biological machines. They start as a simple chain of building blocks (amino acids), then twist and fold into a precise 3D shape. That final shape is everything: it determines what the protein can do, what it can bind to, and how it behaves in the body.
Folding@home uses distributed computing to simulate these motions at massive scale. When you donate compute, you help scientists run more simulations, explore more hypotheses, and accelerate research that would otherwise take far longer.
First: what is a protein?
A simple, intuitive mental model that sticks.
Think “beads on a string.”
A protein starts as a chain of amino acids — like a long necklace. But unlike jewelry, this chain is constantly moving, bumping into itself, and exploring shapes.
Folding is the “click” into function.
The chain settles into shapes that are energetically stable — like a crumpled piece of paper finding a final rest. The final 3D form controls how the protein behaves: what it binds to, what it catalyzes, and what it signals.
Why drugs care: pockets and interactions.
Many medicines work by fitting into a “pocket” on a protein (like a key into a lock). Simulations help reveal where those pockets form, how they open and close, and how a molecule might bind.
So why do we need compute?
Because atoms move… a lot.
Simulations are made of tiny time-slices.
To model protein motion, researchers compute the forces on atoms, then advance the system by a tiny time step… and do it again. And again. Millions (or billions) of times.
Why GPUs are so powerful for this.
Many simulation calculations are “the same math repeated across lots of atoms.” GPUs are designed for massively parallel work like that — which is why modern GPUs can contribute dramatically more folding throughput than CPU-only systems.