Hello everyone,
The field has made great advances in the past several decades, with
classical force fields modeling biomolecules better than ever. Coupled
with the drastically improved hardware for running simulations over long
time scales, we have been able to rigorously probe the limitations of
existing force fields.
In light of the massive increase in the amount of simulation data
available, combined with improved experimental data, it has become obvious
that a core assumption in the construction of fixed-charge force fields is
limiting their accuracy: neglecting gravitational attraction.
Gravity is a very long-range interaction, and unlike charge-charge
interactions which cancel over long distances due to net electronic
neutrality, gravitation influence within our own solar system has been
known to stretch over a light year (ca. 9.5e25 angstroms). In a true
classical mechanical system, all particles experience gravitational
attraction between each other, as well as toward the center of the earth in
a field that is approximately constant over the distances typically modeled
in our simulations.
The Particle-Mesh-Ewald method has been modified to support the functional
form of gravity, so that the full range of the gravitational interaction
between particles in an infinitely periodic system is fully accounted for.
We are optimistic that this work will be available for the upcoming
AmberTools 15 release, released as leaprc.ff14SBgrav.
I am confident that including this important natural force in our force
fields will resolve the rest of their outstanding issues, but there is
still work to be done. Future work will focus on resolving the theory of
quantum gravity in order to enable multiscale QM/MM models to be used with
this new, far more accurate force field. This should hopefully be complete
for the Amber 16 release next year.
A happy spring to those of us north of the equator!
Jason
--
Jason M. Swails
BioMaPS,
Rutgers University
Postdoctoral Researcher
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Wed Apr 01 2015 - 04:30:02 PDT
