Molecular Simulation and Systems Biology - Systems Biology: Networks, Models, and Applications - page 92

will then most certainly involve the development and validation of
a systematic methodology to produce the parameters needed for a
polarizable force field. Very few such force fields exist today, and
generally they have not been very well characterized with respect to
their ability to improve the accuracy of biomolecular modeling
and simulation studies. For important classes of biologically impor-
tant molecules, such as proteins, fixed charge force fields have been
improved over a period of decades. Therefore, in spite of their greater
computational cost, in some respects the current generation of polari-
zable force fields may even be worse than the current generation of
fixed charge models.
There will always be interest in the simulation and modeling of
very large biological structures, many of which are too large for
an atomistic simulation. Therefore, the future will involve consider-
able effort in the development of theoretical approaches to support
coarse-grained and multiscale modeling. In coarse-grained models,
groups of atoms are treated as a single unit and in the simulation
these units interact with an effective potential that can be obtained
in principle from atomistic modeling. In multiscale models, some
regions of a molecular system are treated atomistically (or even
quantum mechanically) and these regions are placed in contact with
others treated with much less detail, such as with a continuum or
finite element model of the material. Currently several research
groups are exploring these areas and a great deal of theoretical work
has yet to be done. However, this work is very important, since there
will always be technologically and medically important biological
phenomena that happen on time and length scales that defy a detailed
atomistic description.
Coarse graining and multiscale approaches allow the modeling
of larger molecular systems. However, there are also small as well as
large molecular systems that exhibit extremely long time scale phe-
nomena. Theoretical advances to be developed in the near future will
certainly include methods to treat the kinetic modeling of rare events.
Some of these will certainly include development and application of
transition path sampling [71] and Markov modeling [72-75].
Benefits and Trends
Improvements in the areas mentioned above will permit qualitative
improvements in our understanding of biology. As stated by McCammon
and Harvey, with each improvement we can perform a larger number
of simulations at a given level of complexity as well as simulate more
complicated systems. The ability to model larger, more complex molec-
ular systems and for longer time scales will help us understand entirely
new types of biophysical processes and phenomena that cannot be
addressed today.