[gmx-users] Strange problem with simple FEP (bug?)

[Maik Goette]

Thanks, but that's indeed not the problem.
I'm just doing a PR-run in state A. State B is not of interest in that case.
Somehow GROMACS mixes things up internally, I guess. The bugzilla-entry 
has been placed.

Regards

Maik Goette, Dipl. Biol.
Max Planck Institute for Biophysical Chemistry
Theoretical & computational biophysics department
Am Fassberg 11
37077 Goettingen
Germany
Tel.  : ++49 551 201 2310
Fax   : ++49 551 201 2302
Email : mgoette[at]mpi-bpc.mpg.de
         mgoette2[at]gwdg.de
WWW   : http://www.mpibpc.gwdg.de/groups/grubmueller/


David Mobley wrote:
> Maik,
> 
>> I'm trying to do a simple FEP within a simple protein, which seems to
>> make things simple...but as you may expect...it is anything else than 
>> that.
> 
>> What I'm trying to do is morphing a Tyrosine into a Phenylalanine in 
>> OPLSAA.
>> Therefore the CZ is changed from the type of TYR to the type of PHE.
>> The oxygen(TYR) is changed to a proton(PHE) and the proton(TYR) to a
>> dummy (PHE). So this is very simple. We checked the tpr-dump and
>> everything looks fine (except, maybe, we missed it). Now, in the
>> position-restraint run (where state B should not be regarded by the
>> system), the OH-proton moves on top of the OH-oxygen and the simulation
>> crashes after a while. We then performed a FEP from TYR to TYR (so A-
>> and B-state are the same) and the system runs.
> 
> I am not an OPLS guy myself, and I don't know the details of your
> transformation, so I'll just tell you one possibility, and you'll have
> to decide whether it could apply in your situation.
> 
> Sometimes, if you are disappearing a Lennard-Jones site that still has
> some amount of electrostatics, it can begin to interpenetrate with
> other atoms which either do not have Lennard-Jones interactions
> (hydrogens in some force fields) or which do (depending on the
> combination rule you use). Obviously, this is bad because (for
> example) protons can move on top of oxygens and cause blowing up.
> 
> A classic example of this is, for example, if I am turning off the LJ
> on an oxygen site at the same time as turning off the charges on that
> oxygen. In many FF's, water hydrogens lack LJ interactions, so they
> can come and overlap with the oxygen I'm disappearing, which is still
> charged (at least to some degree) and cause a "fusion" type event.
> 
> I don't know if that's what's going on for you, but you might think
> about it -- do you have any atoms that you're disappearing, and are
> there other atoms which the combination rules would allow to overlap
> with those?
> 
> The way I get around this problem in my calculations is to always turn
> off electrostatics on any atom I'm disappearing prior to modifying the
> LJ parameters, so that I never have a charged atom which is being
> treated using the "soft-core" potentials that allow overlap.
> 
> Not sure if that's your problem -- just a thought.
> 
> David
> 
> 
>> The whole thing was done with GROMACS 3.3.1 and TIP4P. Any suggestions?
>> Gerrit and me think of a bug somewhere.
>>
>> Regards
>>
>> -- 
>> Maik Goette, Dipl. Biol.
>> Max Planck Institute for Biophysical Chemistry
>> Theoretical & computational biophysics department
>> Am Fassberg 11
>> 37077 Goettingen
>> Germany
>> Tel.  : ++49 551 201 2310
>> Fax   : ++49 551 201 2302
>> Email : mgoette[at]mpi-bpc.mpg.de
>>          mgoette2[at]gwdg.de
>> WWW   : http://www.mpibpc.gwdg.de/groups/grubmueller/
>> _______________________________________________
>> gmx-users mailing list    gmx-users at gromacs.org
>> http://www.gromacs.org/mailman/listinfo/gmx-users
>> Please don't post (un)subscribe requests to the list. Use the
>> www interface or send it to gmx-users-request at gromacs.org.
>> Can't post? Read http://www.gromacs.org/mailing_lists/users.php
>>
> _______________________________________________
> gmx-users mailing list    gmx-users at gromacs.org
> http://www.gromacs.org/mailman/listinfo/gmx-users
> Please don't post (un)subscribe requests to the list. Use the www 
> interface or send it to gmx-users-request at gromacs.org.
> Can't post? Read http://www.gromacs.org/mailing_lists/users.php
> 
> .
>