Re: [AMBER] FF14SB implementation

From: Miroslav Krepl <krepl.seznam.cz>
Date: Tue, 14 Jun 2016 15:14:23 +0200 (CEST)

Dear Vlad,

I have a similar feeling.

As for the ff12SB, the parameters are different there and are not using the
2C atom type:

frcmod.ff12SB:

CX-CT-CA-CA   1    0.126       -21.2            -4.    Phe Tyr chi2
CX-CT-CA-CA   1    0.459        -2.9             2.

frcmod.ff14SB:

CX-CT-CA-CA   1    0.048       180.0            -4.
CX-CT-CA-CA   1    0.000         0.0            -3.
CX-CT-CA-CA   1    0.069       180.0            -2.
CX-CT-CA-CA   1    0.000         0.0             1.

As a result, the ff12SB is OK even without the 2C type.

---------

Another potential issue I see is that the parameters for Phe and Tyr chi2
dihedral seem to be missing from the final table in the Maier et al. 2014
article (Table S8: ff14SB parameters)?
They are discussed in the article but I think they are not written there at
the end. But I am not entirely sure of that. Perhaps I just overlooked them.

Since the ff14SB was already in AMBER some time before the article, I guess
this is unrelated.

Best wishes,

Miroslav



---------- Původní zpráva ----------
Od: Vlad Cojocaru <vlad.cojocaru.mpi-muenster.mpg.de>
Komu: amber.ambermd.org
Datum: 14. 6. 2016 14:55:42
Předmět: Re: [AMBER] FF14SB implementation

"Well, it seems to me that in the aminoct12.lib file which is loaded by
leaprc.ff14SB, the CB atom of PHE and TYR has a CT type and not 2C as
should be based on the paper (Maier et al, supporting info). Unless I am
missing something, this is a very unfortunate mistake that was not
spotted by anybody till now although ff14SB is in use for a while now.
The more puzzling thing is the difference to ff12SB because the
leaprc.ff12SB loads exactly the same libraries, so there shouldn't be
differences between ff12SB and ff14SB (they both load CB in TYR and PHE
as CT type)...

If I am not wrong and the error is in the libraries, it brings up the
importance of some sort of automated check of the topologies created.
Something like a log file with all parameters per residue type present
in the created topology in an easy-to-read format that can be consulted
after the creation of a topology file ... I know parmed has the
functionality to check topologies but still requires users to define
which parameters they would like to check .... Moreover, some automatic
checks with warnings could be useful when new versions of ff are
implemented. .. For example atoms types 2C, 3C, are specific for ff14SB
...Upon creating a topology file, there could be an automated check
script for these atom types and if they are not present where they
should be, a warning could be issued ....

Best
Vlad


On 06/14/2016 02:04 PM, David A Case wrote:
> On Fri, Jun 10, 2016, Miroslav Krepl wrote:
>> I have a question regarding the ff14SB protein ff (the recommended
>> default) as implemented in AMBER. In some of my systems, it
>> seems to me that I am getting better results for tyrosine and
phenylalanine
>> using the preliminary ff12SB
>> version.
>>
>> It almost looks to me as if the new atom type describing the
>> phenylalanine and tyrosine CB atom (type 2C) was not implemented in the
>> residue files at all. There is discrepancy between the AMBER
implementation
>> and the original literature describing the ff14SB.
>>
>>
>> As a result, I observe in simulations that the phenylalanine and
>> tyrosine behave in ff14SB similarly as with ff99SB. Their chi2 dihedral
>> profiles appear to be similar. The same issue is not present with the
>> ff12SB force field version which works differently and in my systems
>> better.
> Thanks for the report. We are looking into the origins of the behavior you
> report, and will post something here as soon as we can. This email is just
a
> place-holder to indicate that we have not forgotten about your mail.
>
> ....regards...dac
>
>
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
>

-- 
Dr. Vlad Cojocaru
Computational Structural Biology Laboratory
Department of Cell and Developmental Biology
Max Planck Institute for Molecular Biomedicine
Röntgenstrasse 20, 48149 Münster, Germany
Tel: +49-251-70365-324; Fax: +49-251-70365-399
Email: vlad.cojocaru[at]mpi-muenster.mpg.de
http://www.mpi-muenster.mpg.de/43241/cojocaru
_______________________________________________
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
Received on Tue Jun 14 2016 - 06:30:02 PDT
Custom Search