Dear John,
I am just adding the reply to this list email so that people can see what I ended up doing.
Yes I did end up figuring out how to use the Magnesium parameters. I can see Tom already answered to point and is updating his files, but I've attached a sample tleap script which does what he is suggesting too.
There are two sets of parameters that you can use.
1) The Aqvist parameters from 1990-ish, which are included in the AMBER force field, or
2) you can the 2007 dummy atom version MD(6)2+ from Oelschlaeger et al. 2007 JMB 366, 687-701.
For option 1)
From emails from Prof. Case, do this:
Go to $AMBERHOME/dat/leap/lib, and edit ions08.cmd
Change the line
i = createAtom Mg+ Mg+ 2.0
to
i = createAtom Mg+ MG 2.0
The type "tleap -f ions08.cmd"
The tleap script expects your Magnesium atom in the pdb file to be named like this:
HETATM 2123 Mg+ Mg+ A 502 45.408 19.600 12.379 1.00 5.17 MG
For option 2)
I've attached files for the 2nd option here, I got them from: http://atlas.physbio.mssm.edu/~tom/.
There are the two parameter files (a frcmod and an off file) and an example tleap script.
The tleap script expects your Magnesium atom in the pdb file to be named like this:
HETATM 2123 MD MG A 502 45.408 19.600 12.379 1.00 5.17 MG
I don't know what MD-engine you plan on using, but a standard version of NAMD will not run with the dummy atoms as it thinks they are hydrogen atoms bonded together and crashes. There is a fix for this in NAMD and I can help you there if you need it, but it does work without a hitch in PMEMD as far as I know.
I hope this helps.
Best,
Jesper
On Feb 20, 2012, at 7:32 AM, Cannon, John F. wrote:
> Dear Jesper,
>
> I was wondering if you ever resolved the problem with the LEaP errors for building the cationic dummy atom model for magnesium (MD6)?
>
> I have tried many variations and get the same errors as you reported. I have also tried cobalt hexamine and get the same type of error.
>
> John Cannon
> Genetics Program Chair and
> Associate Professor of
> Molecular Microbiology and Immunology
> University of Missouri
> 1 Hospital Drive
> Columbia, Missouri 65212
On Feb 20, 2012, at 10:21 AM, Tom Joseph wrote:
> Ah, this is my mistake - I had forgotten to include the lines that inform
> leap of the hybridization types of these dummy atoms:
>
> addAtomTypes {
> { "MM" "" "sp3" }
> { "DX" "" "sp3" }
> { "DY" "" "sp3" }
> { "DZ" "" "sp3" }
> }
>
> Using this leap command your example now works for me (after ensuring all
> the atom and type names are consistent).
>
> Thanks for bringing this to my attention. I've updated the tarball on my
> site to reflect this.
>
> --Tom
>
> 2012/2/20 Cannon, John F. <CannonJ.health.missouri.edu>
>
>> Dear Amber users and developers,
>>
>> I am having a problem that I see in the archives was frequently asked but
>> I see no solutions posted.
>> I am trying to build a protein model with LEaP that has manganese treated
>> with a cationic dummy atom model (MN6).
>> I constantly get a fatal error like:
>>
>>> pp1=loadpdb test.pdb
>> Loading PDB file: ./test.pdb
>> ERROR: Comparing atoms
>> .R<MN6 400>.A<DY2 4>,
>> .R<MN6 400>.A<DY1 3>,
>> .R<MN6 400>.A<DX2 2>, and
>> .R<MN6 400>.A<DX1 1>
>> to atoms
>> .R<MN6 400>.A<DY2 4>,
>> .R<MN6 400>.A<DY1 3>,
>> .R<MN6 400>.A<DZ1 5>, and
>> .R<MN6 400>.A<DX1 1>
>> This error may be due to faulty Connection atoms.
>> !FATAL ERROR----------------------------------------
>> !FATAL: In file [chirality.c], line 140
>> !FATAL: Message: Atom named DZ1 from MN6 did not match !
>> The test.pdb was trimmed down to just
>> HETATM 4722 MX MN6 400 82.265 50.290 15.843 2.0000 1.500
>> M
>> TER
>> END
>>
>> I used Tom Joseph's MD6 and the cobalt hexamine parameters and the same
>> thing happens. I am using Amber 10 and 11 on Linux.
>> When I build the MN6 using createAtom, createResidue, add, bond,
>> createUnit, and zMatrix everything looks good until that loadpdb command.
>> In fact if I omit the last ligand, the error does not occur!
>>
>>
>> John Cannon
>> Genetics Program Chair and
>> Associate Professor of
>> Molecular Microbiology and Immunology
>> University of Missouri
>> 1 Hospital Drive
>> Columbia, Missouri 65212
>> _______________________________________________
>> 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
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
- application/octet-stream attachment: md6.off
Received on Mon Feb 20 2012 - 11:00:03 PST
