Hi Chetna,It may be possible to do this in a tiered approach, where the immediate vicinity (<7A) of the ligand can be made fully flexible, while the backbone of the rest be kept rigid using distance restraints (like an elastic network) with a weak force constant (e.g. 1.5-2 kcal/mol). Then you can remove the part of the protein that is beyond 10A from the ligand, which is expected to contribute less in direct ligand interaction except some long range electrostatic interactions. This will save computational time by avoiding simulating all the 1000 residues.
Wherever there's a break in the chain due to truncation, you have to manually insert TER cards in the input structure, so that tleap knows about the breaks.
Whether this approach is reasonable would depend on the answers you are trying to obtain from the simulations. If you are performing long timescale simulations and expecting to simulate large domain motions in the protein, this most certainly wouldn't work. On the other hand, if you are running short simulations to sample the ligand and side-chain conformations inside the binding cavity and estimate ligand binding free energy, this may work, especially if you don't care about the interaction of solvent with the protein surface far away from the binding cavity.
However any unphysical alteration of the protein comes with potential artifacts even if the alterations are far away from the region of interest (in this case ligand cavity) and they are more probable in longer simulations than short ones. Its important to be aware of that I think.
Best,Supriyo
On Friday, June 28, 2019, 5:34:56 AM PDT, Chetna Tyagi <cheta231.gmail.com> wrote:
Dear Amber users,
I work with a big protein with roughly 1000 residues. The PDB structure has
missing residues which causes problem during equilibration.
Coordinate resetting (SHAKE) cannot be accomplished,
deviation is too large
NITER, NIT, LL, I and J are : 0 0 5803 11719 11720
This region is not useful for our study and I would like it to not be
involved. Sort of
like constraining it while only the ligand-bound domain undergoes
simulation.
Is there a way to do that which will also reduce the computational burden
and time?
If it is possible will there still be such errors? And should I just build
it and re-start?
Thanks--
Chetna
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Sun Jun 30 2019 - 12:00:02 PDT
