[gmx-users] Ions and charge groups

[Guido Polles]

Hi, thanks for the very fast reply.

I used pdb2gmx just to build the protein. Then i wrote my own topology
including the itp file for the protein and for the atp molecule. I
added the ions then.
>From a pdb file I solvated proteon(6 proteins)+atp+ions and added the
water to the top file, that came to look something like

[ system ]
; Name
system in water

[ molecules ]
; Compound     #mols
protein                6
atp                     1
CL                     48
K                      16
SOL                 53231

This way, I get that weird behavior about bonds for CL atoms.
Doing a couple of further tests I actually solved the problem
inverting water and ions both in coordinates and topology, i.e.
[ molecules ]
; Compound     #mols
protein                6
atp                     1
SOL                 53231
CL                     48
K                      16

Everything now is fine, I successfully minimized and running a short
simulation doesn't complain about anything.
I think is a pretty strange behavior and I don't understand why I'm
forced to put ions after water.
Looks like a little bug to me, but I'm not expert enough to be sure.

Anyway, thank you a lot for your help.

Guido




Guido Polles wrote:
> Hi everybody,
>
> Sorry about this, but I'm pretty new to gromacs and I couldn't find
> any clue about my problem in the archives or in the manual.
> I'm using gromacs 4.5.1, and my system is a rather big system (about
> 12x12x12) that includes a proteon, an atp molecule, water and some
> ions.
> I am using ffamber99sb force field. Now, I post the log for my mdrun.
>
>   Initializing Domain Decomposition on 2 nodes
>   Dynamic load balancing: auto
>   Will sort the charge groups at every domain (re)decomposition
>   Initial maximum inter charge-group distances:
>      two-body bonded interactions: 14.813 nm, LJ-14, atoms 25682 25685
>      multi-body bonded interactions: 14.813 nm, Angle, atoms 25685 25686
>   Minimum cell size due to bonded interactions: 16.294 nm
>   Using 0 separate PME nodes
>   Scaling the initial minimum size with 1/0.8 (option -dds) = 1.25
>   Optimizing the DD grid for 2 cells with a minimum initial size of 20.368 nm
>   The maximum allowed number of cells is: X 0 Y 0 Z 0
>
> -------------------------------------------------------
> Program mdrun, VERSION 4.5.1
> Source code file: domdec.c, line: 6428
>
> Fatal error:
> There is no domain decomposition for 2 nodes that is compatible with
> the given box and a minimum cell size of 20.368 nm
> Change the number of nodes or mdrun option -rdd or -dds
> Look in the log file for details on the domain decomposition
> For more information and tips for troubleshooting, please check the GROMACS
> website at http://www.gromacs.org/Documentation/Errors
>
> Of course the main point here is this
>
>   Initial maximum inter charge-group distances:
>       two-body bonded interactions: 14.813 nm, LJ-14, atoms 25682 25685
>       multi-body bonded interactions: 14.813 nm, Angle, atoms 25685 25686
>
> but they're CL ions! Why should they have any bonded interaction or
> angle interaction??

They shouldn't.  Check your [bonds] and [angles] directives for these atom
numbers.  Apparently mdrun is finding bonds to these ions in your topology.
Were these ions added with genion?  Present in the initial structure processed
by pdb2gmx?  Manually added in some location?

> Actually, I made some tests, and:
>
> 1- if I solvate the protein+atp+ions  (I want the ions to be in
> specific positions) system I get that 14nm bond error when I run
> mdrun, but grompp let it go.

OK, so you want ions in specific positions.  Did you put them there?  If so, was
it before or after you generated the topology (presumably with pdb2gmx)?

> 2- in this situation, if I move in the gro file the coords for the
> ions after the water, same happens but now grompp throws a warning
> about charge group distance.

That's probably related to the contents of your .mdp file, which you haven't
posted.  It may or may not be related to the problem at hand.

> 3- if I solvate then use genion to include the ions at least the
> simulation starts. Then I get this error "A charge group moved too far
> between two domain decomposition steps", but that is probably related
> to my system
>

That error message indicates your system has become unstable, likely due to
insufficient energy minimization or equilibration, or incorrect .mdp settings.

http://www.gromacs.org/Documentation/Terminology/Blowing_Up

-Justin

> I really don't get what's going on here.
> Thank you in advance for any feedback.
>
> Guido

--
========================================

Justin A. Lemkul
Ph.D. Candidate
ICTAS Doctoral Scholar
MILES-IGERT Trainee
Department of Biochemistry
Virginia Tech
Blacksburg, VA
jalemkul[at]vt.edu | (540) 231-9080
http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin

========================================


------------------------------

Message: 3
Date: Tue, 30 Nov 2010 14:10:05 -0500
From: "Justin A. Lemkul" <jalemkul at vt.edu>
Subject: Re: [gmx-users] Ions and charge groups
To: Discussion list for GROMACS users <gmx-users at gromacs.org>
Message-ID: <4CF54C0D.9070001 at vt.edu>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed



Justin A. Lemkul wrote:
>
>
> Guido Polles wrote:
>> Hi everybody,
>>
>> Sorry about this, but I'm pretty new to gromacs and I couldn't find
>> any clue about my problem in the archives or in the manual.
>> I'm using gromacs 4.5.1, and my system is a rather big system (about
>> 12x12x12) that includes a proteon, an atp molecule, water and some
>> ions.
>> I am using ffamber99sb force field. Now, I post the log for my mdrun.
>>
>>   Initializing Domain Decomposition on 2 nodes
>>   Dynamic load balancing: auto
>>   Will sort the charge groups at every domain (re)decomposition
>>   Initial maximum inter charge-group distances:
>>      two-body bonded interactions: 14.813 nm, LJ-14, atoms 25682 25685
>>      multi-body bonded interactions: 14.813 nm, Angle, atoms 25685 25686
>>   Minimum cell size due to bonded interactions: 16.294 nm
>>   Using 0 separate PME nodes
>>   Scaling the initial minimum size with 1/0.8 (option -dds) = 1.25
>>   Optimizing the DD grid for 2 cells with a minimum initial size of
>> 20.368 nm
>>   The maximum allowed number of cells is: X 0 Y 0 Z 0
>>
>> -------------------------------------------------------
>> Program mdrun, VERSION 4.5.1
>> Source code file: domdec.c, line: 6428
>>
>> Fatal error:
>> There is no domain decomposition for 2 nodes that is compatible with
>> the given box and a minimum cell size of 20.368 nm
>> Change the number of nodes or mdrun option -rdd or -dds
>> Look in the log file for details on the domain decomposition
>> For more information and tips for troubleshooting, please check the
>> GROMACS
>> website at http://www.gromacs.org/Documentation/Errors
>>
>> Of course the main point here is this
>>
>>   Initial maximum inter charge-group distances:
>>       two-body bonded interactions: 14.813 nm, LJ-14, atoms 25682 25685
>>       multi-body bonded interactions: 14.813 nm, Angle, atoms 25685 25686
>>
>> but they're CL ions! Why should they have any bonded interaction or
>> angle interaction??
>
> They shouldn't.  Check your [bonds] and [angles] directives for these
> atom numbers.  Apparently mdrun is finding bonds to these ions in your
> topology. Were these ions added with genion?  Present in the initial
> structure processed by pdb2gmx?  Manually added in some location?
>
>> Actually, I made some tests, and:
>>
>> 1- if I solvate the protein+atp+ions  (I want the ions to be in
>> specific positions) system I get that 14nm bond error when I run
>> mdrun, but grompp let it go.
>
> OK, so you want ions in specific positions.  Did you put them there?  If
> so, was it before or after you generated the topology (presumably with
> pdb2gmx)?
>

Also relevant here, but I forgot to ask: are you using distance restraints to
keep these ions in place?  If so, that's your problem.  If the restraints can't
be placed in the same DD cell, you have to use particle decomposition (mdrun
-pd), which is somewhat slower.

-Justin

--
========================================

Justin A. Lemkul
Ph.D. Candidate
ICTAS Doctoral Scholar
MILES-IGERT Trainee
Department of Biochemistry
Virginia Tech
Blacksburg, VA
jalemkul[at]vt.edu | (540) 231-9080
http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin

========================================