AMBER Archive (2009)

Subject: Re: [AMBER] Inquiries about antechamber

From: FyD (fyd_at_q4md-forcefieldtools.org)
Date: Thu Aug 27 2009 - 01:15:55 CDT

Dear Yikan,

> Thank you for your reply!
> I think TMS residue shuold be a part of the nucleic acid and I have
> already changed all of these HETATOM into ATOM in 1X6W_3.pdb. Also, there
> was 2 models in 1X6W.pdb and I deleted one of them in 1X6W_3.pdb. Still that
> 1X6W_3.pdb cannot be recoginized by antechamber.
> Any suggestions?

I looked at your 1X6W_3.pdb structure.

Several remarks:
- the TMS part of your molecule does not follow the PDB file format.
Two space characters are missing before the column of atom numbers.
See the PDB file format @ http://www.wwpdb.org/docs.html

- in your molecule you have two parts: (1) the "knwon" nucleic acid
part which is already parametrized/"known" (in the Amber99SB force
field) and (2) the new/"unknown" part: your TMS nucleotide which is
not yet parametrized and which is covalently connected to the part (1).

- to be recognized by an Amber FF a molecule has to present the same
residue and atom names than these available in the Amber force field
topology database (Amber99SB FFTopDB). You have to carefully check
that for your part (1).

- concerning the part (2), you need to develop your own force field
library, but for a molecular fragment, which will be covalently
connected to the part (1). More generally you use Antechamber or
R.E.D. to parametrize the unknown part of a molecular system; this
means in your case, the part (2). Thus, usually you do not feed
Antechamber or R.E.D. with such a big/"undefined" structure.

Conclusions:
- You need to develop a new force field library for the TMS molecule &
more exactly for a TMS molecular fragment compataible with the
Amber99SB FFTopDB.

- You should use the Amber99SB force field & not GAFF.

- Antechamber is strongly connected to GAFF & does not automatically
derive charges for molecular fragments.

- I suggest you to use R.E.D. http://q4md-forcefieldtools.org/RED/ or
R.E.D. Server http://q4md-forcefieldtools.org/REDS/ & to look at our
tutorials @ http://q4md-forcefieldtools.org/Tutorial/ to build the FF
library you need for this fragment.

I hope this helps.

regards, Francois

> On Thu, Aug 27, 2009 at 5:02 AM, Nicholas Musolino <musolino_at_mit.edu> wrote:
>
>> Dear Yikan,
>>
>> I have a question: Does the TMS residue that you mentioned in your e-mail
>> consist of the HETATOM entries in your PDB? If I am looking at it right,
>> the PDB contains a nucleic acid, and it should be one large,
>> covalently-bonded molecule, correct?
>>
>> Remember, antechamber can handle only one species (covalently bonded
>> molecule) at a time. If your TMS residue should be part of the nucleic
>> acid, why are they listed as HETATOM? I don't know whether antechamber
>> distinguishes between ATOM and HETATOM entries in a PDB (the way CDCC
>> Mercury does, see attached PNG file), but I think this is a bit confusing.
>> Probably a result of the software used to generate the PDB.
>>
>> The other issue is that your PDB file lists two MODELS, with very similar
>> coordinates (see lines that contain "MODEL 1" and "MODEL 2").
>> Maybe you could try cleaning up this PDB file, and then sending it to
>> antechamber?
>>
>> Good luck!
>>
>> Nick Musolino
>> ---------------------------------------------------------------------------
>> Nicholas Musolino
>> Ph.D. candidate, Department of Chemical Engineering, MIT
>> musolino_at_mit.edu | 617-253-6675 | Room E19-528
>>
>>
>> On Aug 26, 2009, at 1:07 PM, Yikan Chen wrote:
>>
>> Hey,
>>> I have downloaded a pdb file from www.rcsb.org, as shown in the
>>> attachment
>>> (1X6W.pdb).
>>> It cannot be regonized by Amber force fields directly because there is a
>>> TMS residue that is not included in any force field. So I use antechamber
>>> to
>>> generate a new pdb file and run it in Gaff. It goes like follows:
>>> $AMBERHOME/exe/antechamber -i 1X6W_3.pdb -fi pdb -o newff.pdb -fo pdb
>>> And 1X6W_3.pdb is revised pdb file for antechamber. There are error
>>> messages like the following:
>>>
>>> Unrecognized atomic name JB, exit
>>> Unrecognized atomic name JB, exit
>>> Unrecognized atomic name JA, exit
>>> For atom[28]:E2, the best APS is not zero, bonds involved by this atom are
>>> frozen
>>> For atom[29]:GA, the best APS is not zero, bonds involved by this atom are
>>> frozen
>>> For atom[30]:GB, the best APS is not zero, bonds involved by this atom are
>>> frozen
>>> For atom[31]:GC, the best APS is not zero, bonds involved by this atom are
>>> frozen
>>> For atom[32]:A1TT, the best APS is not zero, bonds involved by this atom
>>> are
>>> frozen
>>> For atom[33]:B1TT, the best APS is not zero, bonds involved by this atom
>>> are
>>> frozen
>>> For atom[38]:E3, the best APS is not zero, bonds involved by this atom are
>>> frozen
>>> For atom[41]:B6, the best APS is not zero, bonds involved by this atom are
>>> frozen
>>> For atom[42]:B7, the best APS is not zero, bonds involved by this atom are
>>> frozen
>>> For atom[46]:JB, the best APS is not zero, bonds involved by this atom are
>>> frozen
>>> For atom[47]:JA, the best APS is not zero, bonds involved by this atom are
>>> frozen
>>> For atom[48]:KA, the best APS is not zero, bonds involved by this atom are
>>> frozen
>>> For atom[49]:KB, the best APS is not zero, bonds involved by this atom are
>>> frozen
>>> For atom[50]:LB, the best APS is not zero, bonds involved by this atom are
>>> frozen
>>> For atom[51]:LA, the best APS is not zero, bonds involved by this atom are
>>> frozen
>>> ---Judge bond type for Residue 1 with ID of 1 and Name of S ---
>>> The frozen atom type can only be 1, 2, 3, 7 (aromatic single), 8 (aromatic
>>> double)Error: cannot run
>>> "/itc/apps/amber/10/amber10/bin/bin.Linux.x86_64/bondtype -j full -i
>>> ANTECHAMBER_BOND_TYPE.AC0 -o
>>> ANTECHAMBER_BOND_TYPE.AC<http://antechamber_bond_type.ac/>-f ac" in
>>> judgebondtype() of antechamber.c properly, exit
>>>
>>> What is that problem? It really puzzles me!
>>> Thank you very much!
>>>
>>> Best,
>>> Yikan
>>> <1X6W.pdb><1X6W_3.pdb>_______________________________________________

_______________________________________________
AMBER mailing list
AMBER_at_ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber