Hi,
I am using a dock-design protocol to design a binding pocket to bind a ligand. I know where parts of the ligand should bind (orinetations and the position in protein) but the other parts can be flexible. Right now after the dock and design protocol, most of my outcomes change the whole position of ligand and don't keep that part fixed. Is there a way to tell the program to keep parts of the ligand fixed? or a way to generate conformers for only parts of the ligand?
Thanks.
Post Situation:
The conformer generation is handled externally to Rosetta - which conformers are allowed to be sampled are entirely specified by the conformers you give to the molfile_to_params.py script. So if you don't want to sample internal flexibility in a certain sub-part of the ligand, simply omit sampling those torsions when generating your conformers. Or alternatively, use the full conformational generation, and then filter out the conformers which have flexibility in that region.
Generally, though, while the gross rotamer changes aren't sampled, there's typically additional minimization which happens on the ligand. You can either turn this off with options, or you can edit the generated params file to remove the CHI lines which correspond to torsions in the fixed region. If there are no CHI lines for a torsion, Rosetta will not recognize it as a rotatable bond.
The final wrinkle in the whole thing is that when Rosetta substitutes conformers, it does so by superimposing the NBR atom. If you have a fixed core, you want to make sure that the NBR atom is located in this core. Otherwise, while the internal conformation of the core will be fixed, changes to torsions between the core and the NBR atom may make the core swing around in the binding pocket. (As when the torsion changes, the NBR atom will be fixed and the other side of the torsion will move.)
The other small issue is rigid body movement. The typical docking protocol will do both coarse rigid body optimization (Translate/Rotate/SlideTogether/Transform) as well as fine optimization of the rigid body positon (through minimization). If you want to hold the ligand fixed, you will want to turn this sampling off, either through omitting the movers, or by picking options which turn off rigid body minimization for the ligand.