You are here

does rosetta good at peptide-protein docking?

4 posts / 0 new
Last post
does rosetta good at peptide-protein docking?

Generally speak the peptide docking is much more chalenged than protein-protein docking for the peptide backbone is much more flexible than protein. Because of this reason, many peptide docking tools would introduce annealing and solvent model. I am wondering, does Rosetta also good at such kind of work?

If I want to do a peptide docking, may I also use solvent model(such as TIP3 and so on) and annealing during the docking?


Post Situation: 
Mon, 2010-12-27 18:51

FlexPepDock coming in 3.2 is the Rosetta program for peptide docking. It's not quite out yet.

Rosetta as a general rule does not use explicit solvent models. There are a few tools kicking around but mostly we use implicit solvent.

What specifically do you mean by annealing? I assume FlexPepDock uses either a constant temperature or an annealing gradient during its Monte Carlo search but I do not know which.

Mon, 2010-12-27 19:00

FlexPepDock has a free server, you can submit an initial peptide-protein complex and the peptide backbone should be in the protein binding site.

Tue, 2010-12-28 00:04

SunH, thanks for pointing that out!

Generally, the Rosetta servers (Design, Dock, Robetta (abinitio), FlexPepDock, others) do a great job for a first-pass model of a problem. A local installation of Rosetta running on a supercomputer/cluster will give better results than the servers will because a lot more computing muscle is available. (Actually, RosettaDesign can do complete treatments of most fixed-backbone problems provided the design space is sufficiently small - it's a tired, wheezy old machine but the fixed-backbone problem is so well treated by Rosetta that even old hardware can handle it.)

Tue, 2010-12-28 10:56