The best advice I can give is check out the LCPO paper and see how they did
it. It's cited in the Amber manual when discussing gbsa.
All the best,
Jason
On Thu, Apr 7, 2011 at 1:30 AM, Bernhard Poll <poll.chemie.uni-hamburg.de>wrote:
> Dear Jason,
>
> thanks for all the information. Made me a little bit smarter:) So I checked
> the mdread.f (I think that’s the code you mean which is -wow- pretty
> impressive:) ). So maybe a stupid questition but how do I get the
> parameters
> for bromine?
> I'm also doing some of my calculation with gbsa=1 again to compare the
> difference to gbsa=2. In my opinion the difference is acceptable small so
> they are still comparable but my boss is just a little bit squeamish so if
> someone could give me a hint how to get those parameters that would be
> fantastic :)
>
> Thanks
>
> Bernhard
>
>
> -----Ursprüngliche Nachricht-----
> Von: Jason Swails [mailto:jason.swails.gmail.com]
> Gesendet: Mittwoch, 6. April 2011 23:31
> An: AMBER Mailing List
> Betreff: Re: [AMBER] bad atom type
>
> Hello,
>
> gbsa=1 uses LCPO to calculate the surface area (and the gradients wrt to
> the
> surface areas), whereas gbsa=2 uses a different algorithm (it starts from
> an
> icosahedron and recursively calculates spheres around atoms; it's described
> more thoroughly in the manual and even more thoroughly in the code ;) ).
> While they give similar answers, the only way to know for sure whether the
> calculations need to be redone is to, in my opinion, redo the calculations
> to check.
>
> Luckily, though, GBSA calculations are fairly cheap, and re-running
> calculations should be trivial if you've kept input files.
>
> One more comment: the reason that gbsa=2 gave you that error and gbsa=1
> didn't is because the LCPO parameters (p1, p2, p3, and p4) are assumed to
> be
> the same as for a typical Carbon atom if the proper atom type is not found.
>
> The only way to fix this approximation (if it's not good enough for you) is
> to actually calculate the parameters that *should* be used for Br and add
> them to the code.
>
> Hope this helps,
> Jason
>
> On Tue, Apr 5, 2011 at 7:56 AM, Bernhard Poll
> <poll.chemie.uni-hamburg.de>wrote:
>
> > Hi Amber users,
> >
> >
> >
> > I calculated free energies using MMPBSA.py using a complex including
> > bromine. So I got the following error
> >
> >
> >
> > Bad atom type: br
> >
> >
> >
> > Refering to an older mailing list entry I used the make-mdins flag and
> > changed the gbsa from gbsa=2 to gbsa=1 an started the
> >
> > calculation again which worked fine then. I already calculated the
> energies
> > of some other ligands docked to the same protein using gbsa=2 and want to
> > compare
> >
> > those. So my questition is if they are still comparable or how big is the
> > difference between those options gbsa=1 and gbsa=2?If they are not is
> there
> > a faster option to
> >
> > get away from this bad atom type error or do I have to calculate the
> other
> > complexes with gbsa=1 again?
> >
> >
> >
> > Thanks in advance
> >
> >
> >
> > Bernhard Poll
> >
> >
> >
> > _______________________________________________
> > 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
>
--
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
Received on Thu Apr 07 2011 - 08:30:06 PDT
