You are here

homology modelling with DNA

5 posts / 0 new
Last post
homology modelling with DNA

Hi all

I'm wondering if there is a way of doing Homology modelling of a protein complexed with the DNA, is there a way to keep the DNA in the template structure? it would be already great if the DNA could be considered at occluded space.


Post Situation: 
Fri, 2012-09-28 06:35

I assume you already tried this and it didn't work?

I can't find any methods for doing homology modeling with non-protein around. How much DNA do you have? If you want the DNA to just occlude volume, we may be able to hack it in as a nonmoving ligand rather than DNA (not that I'm aware of a way to do homology modeling with ligands, either).

If it's homology modeling, then Rosetta shouldn't be able to insert too much junk into where the DNA should go, so you may be able to just do the homology modeling as-is, then use relax or loop modeling to get it to accept the DNA?

You might also be able to use CoordinateConstraints to define "negative space" where the DNA goes - Define a collection of coordinates inside the DNA, then constrain your problematic loops to be in that density, but then use a reverse-valued function with the constraint to cause the loops to score better outside of that density. This might not work if the loops are near the N-terminus because they may move around too much in space for CoordinateConstraints to be useful.

Tue, 2012-10-09 09:12

The released versions of homology modeling can't really handle DNA, ligands, or other non-protein portions. There are some protocols in development that are designed to take that information into account, but nothing that's been released yet.

After talking with TJ (one of the people working on this), I'd say the recommendation for using the released version is to run the homology modeling protocol without any DNA present, and then do a post-filtering analysis to select those models which are most compatible with the presence of DNA (potentially with relax/remodeling like Steven suggests to make minor adjustments). Additionally, you can use constraints to restrict the protein from folding into a form that would occlude the DNA. (I'm not certain if CoordinateConstraints would work all that well with homology modeling, as my understanding is that during the modeling things are allowed to rotate and translate freely - but you can give it a try and see how things work.)

Wed, 2012-10-10 11:28

I was thinking maybe after having selected the best homology model of my protein i could dock the DNA (even a structural superimposition with the template ) and then I could do a "relax_around_chemically_bound_ligand", but the problem I'm having is in the generation of the *.params for the DNA.
I'm following the tutorial in the 3.4 version.
Firstly I have to create the *.mol and then the *.params for my DNA (ligand) but finally it seems like rosetta doesn't recognize my ligand as a DNA chain.
Do you have any idea how I can trick Rosetta ?


Thu, 2012-10-11 01:31

You shouldn't need to create params file for the DNA - there should be params for the nucleotides in the database already. However, you wouldn't treat it as a single ligand - it would have to be inputted as a polymer (in the input file like a normal DNA chain would be from the PDB). The relax_around_chemically_bound_ligand demo can be used, but you would just skip the "Creating a ligand parameter file" section and the "Using the newly created ligand and modified residue parameter files" sections.

You would just make the constraints and run the relax protocol. Note that for the constraints, you would have to account for the fact that the DNA is being read in as multiple residues. (Each nucleotide would be its own residue.)

Thu, 2012-10-11 12:20