You are here

SymDock for docking only the ligand?

5 posts / 0 new
Last post
SymDock for docking only the ligand?

Hi everybody,

another symmetric docking question. What I want to do is symmetrically dock two ligands to a receptor dimer, the binding sides are between the two symmetric interfaces (its the tryptophan repressor dimer binding two tryptophans). When I try SymDock it docks the monomers, as its supposed to be, albeit not in the right way (the crystal structure). What I basically want is just symmetric ligand docking, it should leave the interface from the .sym file untouched, apart from minor adjustments if necessary. Maybe SymDock is the wrong approach in this context?
What does the "-symmetry:initialize_rigid_body_dofs" flag do? Is it a way to preserve the interface from the .sym file?



Post Situation: 
Mon, 2013-06-17 09:53

some more illustrative details ...
The fig example_original shows the input structure as it is supposed to be, the output from the make .sym file (.symm and .modeAB) look essentially the same. The fig example_result shows one of the SymDock output. All look like this.
Also, find attached the .flags file.


Tue, 2013-06-18 06:56

It looks the reply I thought I posted yesterday didn't go through.

SymDock probably won't work for you - it's intended to do docking across the degrees of freedom that define the symmetry.

Instead I'd look into regular ligand docking through the RosettaScripts interface. (There should be a number of demos.) What you can then do is set up the pose with symmetry prior to starting the ligand docking portion of the protocol. I believe (but have not tried) that the ligand docking protocol on a symmetric pose should do what you want it to. (It will symmetrically modify the ligand-protein interfaces, but not change the protein-protein orientation across the interface.) There may be a little debugging involved, though.

Tue, 2013-06-18 11:29

I will try RosettaScripts and report on that matter.
As a first approach I now tried the normal docking, I still have two chains (both take part in the interface) but it looks as if he involves the second chain in the docking, at least judging from the fact that he also does rearrangement in second chain side-chains in the vicinity of the ligand. Can there still be a problem with the energy function or other more internal matters of the docking when using two chains that both contribute to the interface in classic docking?

Thu, 2013-07-11 02:30

I'm sorry, I'm not understanding how you set things up and what you're seeing versus what you're expecting to see.

If you do regular ligand docking with a "symmetric" input structure, what you'll likely see is that you'll get docking of one ligand molecule against the protein dimer, but with no symmetry imposed on the dimer. (The modifications on one chain will not be mirrored on the other chain.) There would also be no preference as to which chain ("first" or "second") to which the ligand would dock.

If the symmetry is such that the two hypothesized ligand binding sites are well separated from each other, regular ligand docking with a "nonsymmetric" homodimer might be the way to go. You can get the binding site of one half, and then manually transfer that binding conformation to the other symmetric half.

Thu, 2013-07-11 11:43