On Fri, Aug 1, 2014 at 5:17 PM, bharat lakhani <lakhbharat.gmail.com> wrote:
> I am running MM-PBSA ALA scanning module. Final output contains this
> following error.
>
> WARNING: INCONSISTENCIES EXIST WITHIN INTERNAL POTENTIAL
> TERMS. THE VALIDITY OF THESE RESULTS ARE HIGHLY QUESTIONABLE
> Complex:
>
This warning is printed when you are using a single trajectory of the bound
complex to extract receptor and ligand conformations. When this is the
case, the internal potential terms (bond, angle, dihedral, 1-4
electrostatics, and 1-4 van der Waals) interactions should all be zero upon
binding. This is true because the conformations are _exactly_ the same in
the bound and the unbound state, so the energies arising between atoms
bonded to each other should be zero.
I can only think of two things that would make this false:
1. You used a different force field for the systems in the bound vs. the
unbound state.
2. You are applying MM/PBSA to a covalently bound complex and you have
broken at least one bond between the bound and unbound states.
Neither of the above are appropriate for MM/PBSA studies.
Since the differences in bond, angle, torsion, and 1-4 nonbonded energies
are identical in your mutant and wild-type systems, I suspect you have
broken a bond. In this case, you must either think very carefully about
the meaning behind the numbers you got to derive physical insight from your
results [1] or choose a different method more suited for studying real
chemistry, like QM/MM methodologies. [2]
If the problem is the first that I mentioned, make sure you use consistent
force fields (ante-MMPBSA.py can help here).
HTH,
Jason
[1] Keep in mind that your binding free energies from MM/PBSA are entirely
without meaning when you have broken a bond -- very few force fields are
parametrized for chemical reactivity. You will need strong justification
(and probably quantification) for the cancellation of errors if you want to
compare the wild type and mutant results.
[2] By "real chemistry" I mean bonds are breaking and/or forming.
--
Jason M. Swails
BioMaPS,
Rutgers University
Postdoctoral Researcher
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Fri Aug 01 2014 - 15:00:02 PDT
