[gmx-users] Fwd: what is the force function for proper dihedrals

[David van der Spoel]

chris.neale at utoronto.ca wrote:
> Haven't looked into the details yet, but this is worth noting for those 
> interested. Thanks William.
> 
> ----- Forwarded message from wnoid at hec.utah.edu -----
>     Date: Sat, 01 Sep 2007 00:15:25 -0400
>     From: William Noid <wnoid at hec.utah.edu>
> Reply-To: William Noid <wnoid at hec.utah.edu>
>  Subject: what is the force function for proper dihedrals
>       To: chris.neale at utoronto.ca
> 
> howdy,
> 
> 
> i am guessing that you want a formula for the cartesian force on the
> atoms involved in a 4-body bonded interaction that is parameterized by a
> dihedral angle.  if so then all you have to do (of course) is to work
> out a sort of nasty jacobian transforming the coordinates, but of course
> it is kind of messy/cumbersome and easy to make a mistake.  the only
> published place i know where they have the answer explicitly is in the
> dl_poly manual, which you can find at
> http://www.cse.scitech.ac.uk/ccg/software/DL_POLY/.  they work it out in
> gory detail on page 18 of the manual (which is actually page 30 of the
> pdf file).  i can't promise there are no typos there, but i think it is
> quite likely correct.  i have explicitly checked their calculation for
> valence angles, though this is considerably easier.  but at least this
> would give you something to check against.  one warning in advance: the
> dl_poly folks define vectors in the opposite convention from normal.
> see in the figure that r_ij is the vector from i to j - whereas i would
> have defined r_ij as the vector from j to i.
> 
> 
> anyway, i hope this is helpful.  if this is what you wanted maybe you
> could forward the info to the mailing list.  if not, sorry to cause you
> any bother.
> 

there is a conference proceedings specifying how gromacs does it, which 
is slightly more efficient than computing derivatives straight away. 
looking at the title I get slightly uncertain. I have the book in my 
office will have a look next week.

@inproceedings{Bekker93b,
         author = {H. Bekker and H. J. C. Berendsen and E. J. Dijkstra 
and S. Achterop and R. v. Drunen and D. v. d. Spoel and A. Sij\-bers and 
H. Keegstra and B. Reitsma and M. K. R. Renardus},
         title = {Gromacs Method of Virial Calculation Using a Single Sum},
         booktitle = {Physics Computing 92},
         pages = {257--261},
         year = {1993},
         editor = {R. A. de Groot and J. Nadrchal},
         address = {Singapore},
         publisher = {World Scientific},

}



-- 
David van der Spoel, Ph.D.
Molec. Biophys. group, Dept. of Cell & Molec. Biol., Uppsala University.
Box 596, 75124 Uppsala, Sweden. Phone:	+46184714205. Fax: +4618511755.
spoel at xray.bmc.uu.se	spoel at gromacs.org   http://folding.bmc.uu.se