Just to add to the good comments of Jason since our manuscript on the GB
model development is not yet out:
GB models and radii sets are somewhat independent.
the igb=8 model was actually fit to PB data where both solvent models
used mbondi2 radii, so in principle it will work well with mbondi2. After
fitting the model to PB, we made a few additional tweaks to the radii in
order to improve agreement with explicit solvent data. It is definitely
better to use mbondi3, but if that is not practical our data shows that it
is still better to use igb=8 with mbondi2 than other GB models in amber.
carlos
On Wed, Apr 27, 2011 at 12:57 PM, Jason Swails <jason.swails.gmail.com>wrote:
> Hello,
>
> On Wed, Apr 27, 2011 at 7:48 AM, Oliver Kuhn <oak.amber.googlemail.com
> >wrote:
>
> > Hi Jason,
> > I have read about the igb=8 option.
> > My version of MMPBSA.py tells me that igb=8 is not available.
> > Will it run with AmberTools 1.5?
> >
>
> What Bill said is correct -- this is an orthogonal statement. igb = 8
> currently requires a *different* radius set than what's provided in leap.
> So-called "mbondi3" radius set. It's the mbondi2 radii with modifications
> the carboxylate oxygen radii (in the deprotonated state), and modifications
> to ARG hydrogens.
>
> There is currently no utility available for setting this radius set in the
> Amber distribution, so unless you read the original paper from the
> Simmerling group and assign the radii by hand in the topology file, there's
> no way of *properly* using it right now.
>
> I'm working on a Python script (I do love python...) right now that will
> allow you to change the radii set to any set you want (including the one
> needed for igb = 8), but I'm working out some of the kinks for Nucleic
> acids
> now. We may try to release it afterwards as some sort of supplement to AT
> 1.5
>
>
> > Another question:
> > What do you think about the usage of internal dielectrics 2 and 4 in
> > MMPB/GBSA calcuations?
> > I get results closest to experimental value when using 1.* *But I would
> > also
> > like to know if there is a profound reason to use 1, I mean, is something
> > in
> > the procedure parameterized to 1?
> >
>
> GB should probably use 1, since that was used in the parameterization.
> Note
> that the internal dielectrics have different meanings in GB and PB. In PB,
> it's reasonable to use dielectric constants of 2 or 4 to account for
> electron polarizability (account for electronic degrees of freedom in a
> mean
> field way). When I say reasonable, I mean it's been done before and is
> supported (and probably attacked) in the literature.
>
> The "internal dielectric" for GB, on the other hand, is supposed to be the
> dielectric for vacuum. What GB actually calculates is the delta G of
> solvation in moving from one environment (vacuum) to another (water). The
> internal dielectric constant of the protein itself actually cancels in the
> derivation of the GB equations. Therefore, changing the internal
> dielectric
> constant in GB is akin to changing what you're calculating --> instead of
> looking at the delta G of transferring the molecule from vacuum to water,
> you're transferring it from an organic liquid (of dielectric constant 2 or
> 4
> or whatever you set) into water.
>
> Of course, the method is so parametrized that it's probably hard to take
> much physical meaning out of the dielectric constants. Long story short --
> you should probably leave the GB dielectric constants alone since it's so
> heavily parametrized for water (which is why we removed those as input
> variables in MMPBSA.py), but it's more reasonable to change dielectric
> constants in PB calculations.
>
> HTH,
> Jason
>
>
> >
> > Greets,
> > Oliver
> >
> > --
> > Oliver Kuhn, Department of Bioinformatics,
> > Center for Medical Biotechnology, University of Duisburg-Essen,
> > Universitätsstr. 1-5, 45141 Essen, Germany
> > phone +49 201 183-3121, oliver.kuhn.uni-due.de
> > _______________________________________________
> > AMBER mailing list
> > AMBER.ambermd.org
> > http://lists.ambermd.org/mailman/listinfo/amber
> >
>
>
>
> --
> Jason M. Swails
> Quantum Theory Project,
> University of Florida
> Ph.D. Candidate
> 352-392-4032
> _______________________________________________
> 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 Wed Apr 27 2011 - 11:00:02 PDT
