AMBER Archive (2007)

Subject: AMBER: Internal dielectric and MMPBSA

From: Ileana Stoica (i.stoica_at_ucl.ac.uk)
Date: Tue Jun 12 2007 - 06:32:17 CDT

I am looking at the effect of the internal dielectric constant
on MMPBSA ligand binding energies.

Normally, both the MM Coulomb energies and the PB polar solvation
component should change with the internal dielectric, labelled
DIELC and INDI respectively.

Indeed, if I set DIELC=INDI=1, I get:

# -----------------------
# MEAN STD
# =======================
ELE -32.90 9.16
.....
PBCAL 67.53 7.53

and when I set DIELC=INDI=10, I get as expected:

ELE -3.29 0.92
...
PBCAL 0.56 0.06

However, if in addition to MM=1 and PB=1, I also happen to have
GB=1 in the Mmpbsa input, the results for DIELC=INDI=10 look like:

ELE -32.90 9.16
PBCAL 0.56 0.06

The Coulomb energies are no longer scaled and consequently the new PBTOT is
erroneous.
So, it appears that to get the expected scaling of electrostatics,
we NEED to have in the input file:

PB=1 AND GB=0.

Is this true or am I missing something? It certainly can lead to confusing
results.

Thank you,

Ileana

Ileana Stoica
University College London

-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber_at_scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo_at_scripps.edu