Hi all,
now that traffic on this list quited down a bit - I'd like to run by you
something I've been working on.
I have written Amber/sander interface to APBS which allows access to most
of the APBS functionality from within sander. APBS
(
http://apbs.sourceforge.net/) is a Poisson-Boltzmann solver for continuum
electrostatics with many features. One of the advantages of APBS is that
it can be built as a library and linked with another application for a low
level access to all of APBS functionality.
The actual sander/APBS module is implemented through iAPBS interface (an
APBS library wrapper) and minor additions to the sander code. When the
appropriate keyword is present in the sander input file the apbs_forces
subroutine is called and it returns APBS calculated solvation energies and
forces. These are added to the total energy and forces.
The current implementation uses igb=6 to calculate energies and forces in
vacuum and APBS solvation energies and forces are added to that. So, for
instance, this input file would carry out an implicit solvent
minimization:
&cntrl
maxcyc=200, imin=1,
cut=12.0,
igb=6, ntb=0,
/
&apbs
dime = 33, 33, 33,
cglen = 10.0, 10.0, 10.0,
fglen = 9.0, 9.0, 9.0,
&end
Among currently implemented features are: implicit solvent minimization,
dynamics, simple, single step calculation of solvation energies and
writing out calculated electrostatic properties for visualization using
third party applications (VMD, Pymol, PMV, OpenDX, etc.). Documentation
for the current, beta version of the Amber/APBS module is available at
http://morgan.ucsd.edu/iapbs/ and
http://morgan.ucsd.edu/iapbs/usersguide/index.html The code can be also
donwloaded from there together with Amber9 patches.
So after this long introduction (sorry), I'd like to ask if the outlined
Amber/APBS implementation makes sense. I'm not an Amber expert so although
the current module seems to work fine (in my limited testing) it is
certainly possible I may have overlooked something. One obvious thing is
that the current use of igb=6 is not very clean and it can be probably
confusing for the users. So I'd welcome any comments, recommendations or
ideas you might have.
Thanks!
Robert
--
Robert Konecny
W. M. Keck Laboratory for Integrated Biology
NSF Center for Theoretical Biological Physics
NIH National Biomedical Computation Resource
Center for Research in Biological Systems
9500 Gilman Drive, M/C 0365 E-mail: rok.ucsd.edu
University of California, San Diego Phone: 858-822-5683
San Diego, CA 92092-0365 Fax: 858-534-4974
Received on Wed Apr 12 2006 - 06:07:12 PDT
