I used swiss-model to model my protein of interest (transcription factor) and it gives a monomer as the modelled protein. I want to make a dimer for further analysis.
Have you tried Symmdock software. Its a geometry and symmetry based docking software gives you homodimer out of monomer. You need to give a monomer pdb file and the software will generate homodimer and will send you the results in email. You can also use rosetta symmetric docking server. Try using these.
http://bioinfo3d.cs.tau.ac.il/SymmDock/
http://rosie.rosettacommons.org/symmetric_docking/documentation
In addition to valuable answer from Punam Ghosh, I suggest you trying PDBePISA, a software designed for multimeric state and protein-protein interface exploration. You can check whether your homodimer will be stable or not in solution, along with many other properties.
You can find PDBePISA here: http://www.ebi.ac.uk/pdbe/pisa/
I hope this helps
The best way for generation of any dimer is to use experimental reports, the softwares could do it based on some energy or symmetric parameters, which is far from reality.
I think that all of you are correct :). The experiment gives you the highest confidence on your molecule, but if the template that you've modeled your protein on is very similar (in terms of structural features, also homology, evolution) to your protein, Punam's and Juan's suggestions are very helpful. What I would do in your case would be dependent only on the certainty or trust that I request from my results. Whether you can do it easily and well or you will not be able to do it at all depends on your proteins. You have to assess the risk of wasting time by yourself, or share more details on your proteins and what you want to do with them :) Good Luck!
Thank You very much for all of your sugessions and Grzegorz, I want to model this protein for vertual screening.
For virtual screening you have to have a target with a high resolution of atomic coordinates... I hope that your template is very similar to the modeled protein. I would be rather cautious here. If the identity (not homology) of these proteins is not really high (lets say 80%), I would not trust the model enough to do VS. I know that modeling a protein with 30% and less of homology is not a big challenge sometimes, but the only thing that we can say is, that the overall fold of the protein MIGHT be similar to the one of the template. In such a model exact positions of residues that you want to attach some small molecule to cannot be trusted. All the Best!
- Badges
- Science topic
- Similar topics
- Filing