Re: [AMBER] antechamber limitation on molecule size

From: FyD <fyd.q4md-forcefieldtools.org>
Date: Thu, 08 Nov 2012 10:55:44 +0100

Dear Yi,

> It is just an organic molecule with lots of fused aromatic rings. You can
> see the pdb in attachment. Actually I don't know any methods to reduce it
> into smaller building blocks. If you know how to do this please teach me.
> Thanks a lot.

Your molecule is just... great ;-) Nice project...

We worked on highly similar projects about calixarenes
http://en.wikipedia.org/wiki/Calixarene with or without metal ions.
As usual, see projects in R.E.DD.B. & in particular the 3 following projects:
http://q4md-forcefieldtools.org/REDDB/projects/F-87/
http://q4md-forcefieldtools.org/REDDB/projects/F-88/
http://q4md-forcefieldtools.org/REDDB/projects/F-89/

I definitively think you 'should' split this molecule into different
building blocks; this is straightforward using R.E.D. Server/R.E.D. IV.

You could even compare the two approaches (i) using the whole molecule
(I can push back the 150 atoms limit of R.E.D. Server or you can use
R.E.D. Server Dev.) and (ii) using the building blocks.

About (i), the key is chemical equivalencing (Ante_R.E.D. 1.x from the
R.E.D. tools III.x does NOT handle this case, and our algo. in R.E.D.
Server/Ante_R.E.D. 2.0 is not well adapted to this case. I would be
just curious to see what would be the answer from Antechember here...
We can run the new algo. available in our new R.E.D. version and adapt
the output to create a P2N file for R.E.D. IV if you need it.

About (ii), your just create the elementary building blocks as
originally defined in Cieplak et al.
http://onlinelibrary.wiley.com/doi/10.1002/jcc.540161106/abstract (see
also examples in R.E.DD.B.). Once the R.E.D. Server job is done you
reconstruct your pseudo-macromolecule using a LEaP script; see for
instance:
http://q4md-forcefieldtools.org/REDDB/projects/F-87/script1.ff
The new R.E.D. version will reconstruct each macromolecule by itself;
i.e. without LEaP.
Your big molecule could be split into 4 elementary building blocks;
may be more if you want to create analogs of this big molecule...

(you could also consider adding extra-points & united carbons)

You could write a nice tutorial here, and compare the whole molecule
approach and its limitations versus the building block approach and
its numerous advantages.

regards, Francois


> On Wed, Nov 7, 2012 at 8:34 PM, Aron Broom <broomsday.gmail.com> wrote:
>
>> what is your 150 atom molecule? You're confident there is no reasonable
>> method for reducing it to smaller building blocks?
>>
>> ~Aron
>>
>> On Wed, Nov 7, 2012 at 9:21 PM, <anyiphysics.gmail.com> wrote:
>>
>> > Hi Francois,
>> >
>> > Thanks a lot for your reply. But I still have a question: To derive
>> charges
>> > for central/terminal fragment of a molecule, I need a P2N file first.
>> > However, since my molecule is bigger than 150 atoms, I cannot use
>> ante-RED
>> > to get the P2N file. If I don't have the P2N file, how can I break it
>> into
>> > fragments? If you know the answer please help me. Thank you.
>> >
>> > Yi
>> >
>> > On , FyD <fyd.q4md-forcefieldtools.org> wrote:
>> > > Dear Daniel Roe and Yi An,
>> >
>> >
>> >
>> > > > I think 150 atoms is much larger than anything antechamber was
>> >
>> > > > designed to parameterize off the bat. Something like that you would
>> >
>> > > > typically break down into smaller chemical components and then
>> >
>> > > > parameterize those (analogous to breaking down a peptide into smaller
>> >
>> > > > amino acid chunks instead of trying to parameterize the whole thing).
>> >
>> > > > A good resource for stuff like this is something like RED server
>> >
>> > > > (http://q4md-forcefieldtools.org/REDS/).
>> >
>> >
>> >
>> > > I think one should always consider splitting a large molecule into
>> >
>> > > smaller elementary building blocks and generate the corresponding
>> >
>> > > molecular fragments when one wants to parameterize a large molecule.
>> >
>> > > RED and RED Server are specially designed for this task.
>> >
>> > > See http://q4md-forcefieldtools.org/REDDB/projects/F-90/ for instance.
>> >
>> >
>> >
>> > > RED Server at http://q4md-forcefieldtools.org/REDS/ is limited to
>> >
>> > > 150 atoms.
>> >
>> > > See http://q4md-forcefieldtools.org/REDS/faq.php#17; we can obviously
>> >
>> > > postpone this limit if this is needed by the user; just send a request
>> >
>> > > to contact_at_q4md-forcefieldtools.org.
>> >
>> >
>> >
>> > > RED Server Development at
>> >
>> > > http://q4md-forcefieldtools.org/REDS-Development/faq.php#17 is limited
>> >
>> > > to 350 atoms; This web server handle all the elements of the periodic
>> >
>> > > tables and is mostly designed for metal complexes. It can be obviously
>> >
>> > > used for organic molecules but ambiguities between atom names are not
>> >
>> > > well handled (CA is recognized as a Calcium atom and not as a C alpha
>> >
>> > > carbon). This is handled within the new RED version in preparation.
>> >
>> >
>> >
>> > > regards, Francois
>> >
>> >
>> >
>> >
>> >
>> > > > On Wed, Oct 31, 2012 at 1:05 PM, Yi An anyiphysics.gmail.com> wrote:
>> >
>> > > >> Hi all,
>> >
>> > > >>
>> >
>> > > >> I'm now using antechamber to derive partial charges for some large
>> > > organic
>> >
>> > > >> molecules. My input file is pdb and I want to get mol2 files for
>> these
>> >
>> > > >> molecules. My problem is when the molecule contains more than 150
>> > > atoms, I
>> >
>> > > >> can only get the sqm.out file. I can't get the mol2 file even though
>> > > the
>> >
>> > > >> sqm.out says the geometry is converged. However, if the molecule has
>> > > >> atoms then I don't have this problem.
>> >
>> > > >>
>> >
>> > > >> The method I use in antechamber is AM1 and the commands
>> > > are "antechamber -i
>> >
>> > > >> XXX.pdb -fi pdb -o XXX.mol2 -fo mol2 -c bcc -s 2". Does anyone know
>> if
>> >
>> > > >> antechamber has a limitation on molecule size? Any related
>> information
>> > > is
>> >
>> > > >> appreciated. Thanks.


_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Thu Nov 08 2012 - 02:00:02 PST
Custom Search