is rosetta used to work out the folding of currently unknown protein shapes, or is it still being fine-tuned, with only known protein shapes being tested to see how accurate rosetta is?

You will see both types of work. If the protein's native structure is known, you will see it in the graphic. There will be four boxes with protein images, and the RMSD will be calculated at each step along the way.

If you see only 3 protein images, the native structure is not known and so RMSD, which is a measurement comparing to the actual native structure is unknown as well.

---

Rosetta Moderator: Mod.Sense