I am trying to dock DNA on a protein. I have a pdb file for protein, but for DNA, I downloaded a pdb file from the protein database, which was a complex between a protein from the same family as my protein and DNA with Zn ion. Since I wanted to dock the DNA on my protein, I deleted the protein in that complex and kept only the DNA and Zn atom. Zn is important for the confirmational stability of the protein. Was that a valid approach?
Also, I changed the edesolv weight to 0 for all the steps. Was that necessary?
If those approaches were valid, should I refine the best model generated before MD simulation?
I am using HADDOCK 2.4 web server!
Hi,
keeping the Zn ion is of course a good idea, and using the DNA you found in the PDB makes sense (provided the sequence is the same)
I would not put the weight of the desolvation energy to 0, but rather stick to the optimal parameters for protein-DNA docking (they automatically change if you selected nucleic acids in the dropdown menu at the start).
The web server will refine in water (with a short MD simulation) the top 200 models coming out of the flexible refinement (it1).
You can find an additional tutorial for protein-DNA docking using HADDOCK 2.4 here Protein-DNA docking Using HADDOCK High-Ambiguity Driven DOCKing – Bonvin Lab
1 Like
As a small addition, the desolation energy is indeed better to be turned off for nucleic acids.
This should be what the server does automatically. But can be good to check.
1 Like
Thank you both for the comment! Yeah, the edesolv energy weights were 1.0 for all stages.
So, further water refinement is not necessary after the docking, even if the scores are relatively smaller (less negative)? I have noticed that the scores improved after refinement for the top-ranked cluster.
One more question: If I want to dock the same DNA file with mutations, should I refine the mutated DNA before docking, or is it not necessary?
Thank you both for the comment! Yeah, the edesolv energy weights were 1.0 for all stages.
So, further water refinement is not necessary after the docking, even if the scores are relatively smaller (less negative)?
the currents default is indeed not to do water refinement but only a short EM.
You can of course still do it.
Also regarding the scores, remember the restraint energy is part of it.
You could also run your preferred models through the refinement server (default is water refinement)
This will not have any restraint energy.
Thank you, Dr. Bonvin!
What do you recommend in this case? One more question: If I want to dock the same DNA file with mutations, should I refine the mutated DNA before docking, or is it not necessary?
What do you recommend in this case? One more question: If I want to dock the same DNA file with mutations, should I refine the mutated DNA before docking, or is it not necessary?
No need for refinement I would say.
But the question is rather: What do you want to achieve / test with DNA mutations?
By docking you will always get a complex, i.e. the docking won’t tell you something does not bind.
You can of course compare the scores, but these are not binding affinities / free energies.
Also if you want the test the impact of a few mutations, provided you have a reliable model of the complex, I would rather introduce the mutation in the PDB of the complex and use the refinement interface to get scores (i.e. no full docking)
1 Like
Thank you for prompting that question!
I want to see if the mutations affect the binding and structural conformation.
So just to make sure I understood your suggestion, you recommend introducing mutation in the complex that I get after docking the wild-type DNA sequence and using that to get scores based on the refinement interface?
if that’s correct, could you please tell me how I would use the refinement interface? Thank you!
I want to see if the mutations affect the binding and structural conformation.
So just to make sure I understood your suggestion, you recommend introducing mutation in the complex that I get after docking the wild-type DNA sequence and using that to get scores based on the refinement interface?
This would of course only give you possible differences in the HADDOCK scores and interaction energies.
The binding mode won’t change.
if that’s correct, could you please tell me how I would use the refinement interface? Thank you!
Simply upload a model of the complex or of the mutated one and perform the default water refinement.
That’s it!