[gmx-users] Protein_model_refinement

[Mark Abraham]

On 6/10/2011 3:52 PM, shahid nayeem wrote:
> Dear All
> I want to refine a protein homology model, I did minimization but then 
> procheck does not give better Ramachandran plot.

EM can only do a local minimization, and there is no good reason why 
that should show a marked improvement on an arbitrary homology model.

> This model is a homodimer with each chain of 609 residue and a total 
> of 1218 residue so a long MD simulation to search native state is not 
> feasible. Can any one help me on this list to tell that how should I 
> attempt this problem.

Short MD will also be unlikely to do anything useful, because it is 
strongly determined by the initial conditions. So GROMACS is unlikely to 
be the tool you need for this problem, nor this list the place to find 
it. I would consult the homology modelling literature and software tools.

Mark