Modeling aqueous environments in X-PLOR. How?

Paul H Axelsen axelsen at axe1.med.upenn.edu
Thu Jul 7 12:58:17 EST 1994


gwpoulos at acs4.bu.edu wrote:
: I've been trying to run some molecular dynamics simulations using the
: program X-PLOR.  I'm trying to study a polypeptide with amphipathic
: alpha helices and I need to be able to look at polar and hydrophobic
: interactions.  I was wondering how I might be able to model the
: molecule in an aqueous environment.  Running the dynamics with
: lots of individual water molecules is too computationally expensive,
: since the program tries to also determine all of the dynamics of the
: water molecules.  Any suggestions?  

You've got a problem.  Simulations with hundreds of explicit water molecules
already invoke many severe approximations, but do seem to yield realistic
results for well-conditioned problems.  Various schemes to model water
implicitly inevitably require even more severe approximations, and will
make your results even less realistic (q.v. Axelsen et al, Biophysical 
Journal 54:249-258, 1988).  Within X-plor, the quickest (and dirtiest, but
still very clever) subterfuge is to use a distance-dependent dielectric.

I suggest you find an example of the kind of analysis you want to perform
in the literature, assess the cost in time and resources, and then proceed
as indicated.  X-plor and CHARMM have the same ancestor, but you will find 
a lot more solvent work has been done with CHARMM (or AMBER or GROMOS).

--
--------------------- axe at pharm.med.upenn.edu --------------------------------
                                                                              
Paul H. Axelsen MD            ....   ....  .   .  .   .    215-898-9238 (ofc) 
Department of Pharmacology    .   .  .     ..  .  ..  .    215-898-9766 (lab) 
University of Pennsylvania    ....   ...   . . .  . . .    215-573-2236 (fax) 
36th St. & Hamilton Walk      .      .     .  ..  .  ..    215-293-1659 (home)
Philadelphia, PA 19104-6084   .      ....  .   .  .   .                       

------------------------------------------------------------------------------



More information about the Molmodel mailing list