You are here

Loop modelling before peptide docking

5 posts / 0 new
Last post
Loop modelling before peptide docking

Hi all!

I am using FlexPepDock to dock a peptide to a target. The target has a loop which has missing electron density for 6 residues, and I want to compare how filling in this loop will affect FlexPepDock poses - the loop is too close to the docking position of the peptide for me to be comfortable not dealing with this situation.

It is not clear to me how to proceed. I have the following questions:

1. The overall loop is ~21 amino acids. Presumably, I can just remodel the missing 6 residues. Is this assumption correct?
2. What is the best way to have an initial PDB to feed into the loop remodelling/building protocol. I have tried writing dummy xyz (0.0 0.0 0.0) co-ordinates in a text editor for the six residues, renumbering this modified PDB in pymol (both atom ID and residue number) and then passing this through KIC remodel and refine, but this fails to build the loop with the following error;

protocols.loops.loop_mover.perturb.LoopMover_Perturb_KIC: [WARNING] Failed to build loop with kinematic Mover during initial kinematic perturbation after 100 trials: LOOP 145 149 148 0 1
protocols.loops.loop_mover.perturb.LoopMover_Perturb_KIC: Unable to build this loop - a critical error occured. Moving on ..

3. I am now going to try the "-build_initial" flag using the non-KIC method, but if anyone has a better solution, please do let me know.


Post Situation: 
Tue, 2013-10-01 07:24

I would first build the loop into your PDB via RosettaRemodel. It is included in the newest weekly releases:

Then I would give the PDB to KIC to fully remodel it. If the loop is a full 21 amino acids, you may want to look for potential secondary structural elements. KIC can generate something for 21 aminos, but I wouldn't put much faith into it regardless. You may be better off using CCD with fragments because it is so long..

Tue, 2013-10-01 10:34

Thanks for your response jadolfbr.

I will try RosettaRemodel, but in the meantime, I have now built the initial PDB using the build_initial flag using fragment files and the CCD protocol of loop remodel.

However, it is still not clear to me whether to remodel the full 21 amino acid loop, or just the 6 amino acids that are missing in the crystal structure using the amino acids at either ends as the anchor points.

Any help would be much appreciated.


Wed, 2013-10-02 03:59

Glad you were able to use the build_initial flag. I would probably use the 6 aminos, and maybe one or two on either side. 21 is too much to get anything reasonable and if the crystal was just missing density here, it means you should be able to trust most of those other residues...

Wed, 2013-10-02 09:50

Thanks for your guidance jadolfbr.


Thu, 2013-10-03 08:53