On Thu, 2015-02-26 at 12:46 -0300, George Tzotzos wrote:
> I’m dealing with a homo-dimer (A and B subunits) in complex with two
> identical ligands.
>
> In MMGBSA calculations, I considered as complex the dimer
> (A/ligandA-B/ligandB), as receptor (sub-unit A/ligandA) and as ligand
> (sub-unit B/ligandB). The calculation resulted in the expected output.
>
> I’d like now to determine deltaG binding of ligandA and ligandB. I
> have the parameter files of both ligands. I’m a bit stuck on how to do
> this without re-running MD simulations.
So you have 4 components: ligand A (lA), subunit A (sA), ligand B (lB),
subunit B (sB). You got the binding energy between lA.sA -- lB.sB. What
binding energies are you looking for now? I can think of 5 "popular"
possibilities -- the one you pick depends on what you want to learn:
lA -- sA.lB.sB and lB -- lA.sA.sB ?
lA -- sA.sB and lB -- sA.sB ?
lA -- sA.sB and lB -- lA.sA.sB ?
lB -- sA.sB and lA -- sA.lB.sB ?
lB -- sB and lA -- sA?
or some combination of those? You can either run the simulations of
those systems necessary, or you can generate them from the simulation of
lA.sA.lB.sB that you already did using cpptraj. For example, let's
consider the first two scenarios above:
lA -- sA.lB.sB : you need the original lA.sA.lB.sB topology, the lA
topology, and the sA.lB.sB topology. You can generate the latter two
(ligand and receptor) using ante-MMPBSA.py giving it a set of masks. The
second part of that line is the same idea, just lB instead of lA.
lA -- sA.sB : Your 'complex' topology is now lA.sA.sB, so you need to
use cpptraj to strip lB from the trajectory and print out a compatible
prmtop file along with the corresponding trajectory (you might as well
also strip water and ions so MMPBSA.py doesn't need to do that step).
Then you need to create sA.sB and lA topologies (you can get those from
ante-MMPBSA.py, as well).
That gives you the basic template for how to extract trajectories for
any of the systems above you want to define as the 'complex'... just
note that the more you strip away, the more perturbed your system is and
the less representative the generated trajectory will be of the 'true'
trajectory. At some point (who knows what point), cancellation of
errors won't be enough. But these seem to me to be outstanding,
unanswered questions in the field.
HTH,
Jason
--
Jason M. Swails
BioMaPS,
Rutgers University
Postdoctoral Researcher
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Thu Feb 26 2015 - 08:30:02 PST
