Mag. Thomas Taylor wrote:
> Hi!
>
> According to the AmberTools docs (p35) something like 'ERROR: Flag "TITLE" not
> found in PARM file' means that my SANDER will likely not work with my CHAMBER
> generated prmtop.
> However I have Amber10 and patched it with the latest bugfix.all (as of today,
> 2010-01-12) and keep seeing this error.
SANDER10 does not support the CHARMM enabled prmtop files which chamber
(in AmberTools 1.3) produces; support for these was only backported into
PMEMD10 via bug fix 26:
http://ambermd.org/bugfixes/10.0/bugfix.26
> Is this just a "minor" bug (CTITLE is not processed correctly but changing
> CTITLE to TITLE gives correct results)
This is not a bug, it is a safeguard; we were very worried that a user
would accidentally attempt to use a CHARMM enabled prmtop file with an
earlier version of SANDER or PMEMD. These do not contain the code to
evaluate the extra CHARMM energy terms and as a result, would generate
incorrect results devoid of contributions from these terms.
In our testing (with a CHARMM prmtop file in "TITLE" form), older
versions of PMEMD would run molecular dynamics, but latently give the
wrong forces and energies because the code was not using the correct
underlying potential function. The situation with SANDER10 was similar,
but made a little more complex due to an incomplete preliminary
implementation of CHARMM support accidentally slipping into the SANDER10
release. It would recognize that a CHARMM prmtop file was being used
during the parsing stage and then quit. Again, it lacked the ability to
evaluate the extra energies and forces pertaining to the extra CHARMM terms.
The CTITLE flag was a solution to all of this; essentially it would
cause any older versions of SANDER/PMEMD to fail.
> or is it not possible to use CHAMBER
> (1.3) with AMBER10? (bugfix.26 has me believe otherwise and grepping for
> "CTITLE" revealed an occurence in the patched source-code)
No; the only public AMBER molecular dynamics engine that supports such
chamber prmtops at this moment in time is PMEMD10 via bugfix 26.
SANDER11 will support this when it is released in AMBER11.
Attempting to circumvent this CTITLE protection mechanism by hand and
use such a prmtop file in SANDER10 or PMEMD10 will give you incorrect
energies and forces as outlined. Essentially, I think you would end up
with some form of a freaky Frankenstein projection of the CHARMM
parameters on the basis of an AMBER potential.
--
Mark Williamson, Post Doc
Walker Molecular Dynamics Group
Room 395E
San Diego Supercomputer Center
9500 Gilman Drive
La Jolla, CA 92093-0505
Email: mjw at sdsc.edu
Office: 858-246-0827
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Tue Jan 12 2010 - 18:00:03 PST
