[gmx-developers] Flat-bottom potentials

[Jochen Hub]

Hi developers,

I would (finally) like to implement some flat-bottom quadratic 
potentials into Gromacs. They are useful for many applications, but so 
far I usually implemented them as more or less dirty hacks, which of 
course never made it into the main branch. Berk, David and me talked 
about that topic already years ago, and now I would like to continue 
this debate.

I think the flat-bottom quadratic potentials should allow:

1) different geometries:
   - spherical (e.g. to avoid evaporation from a spherical droplet)
   - cylindrical in x-y plane, but free in z (needed to compute PMFs for 
membrane channels, to keep the solute aligned with the channel pore)
   - 1D, e.g. to keep a layer of some thickness stable

2) different reference positions. This is important if you compute PMFs 
for mutimeric membrane channels, where you want to have a cylindrical 
flat-bottom quadratic potential at each monomer.

I can think of two ways to implement them:
a) not preferred: into the pull code. AFAIK, that would not allow 
different reference positions at the moment (tell me if I am wrong).

b) preferred: New function types (ft) for the [ position_restraints ] 
section in the topology. So ft=1 would be the normal posres, and ft 2, 
3, 4 5, 6 could be for spherical, cylindrical in the xy plane, and 1D in 
X, Y, and Z, respectively. So ft>1 would ask for one additional 
parameter r which gives the radius/half-width of the flat bottom. Only 
one force constant would be required for the flat-bottom potential.

Would you agree with that? Please let me know.

Cheers,
Jochen



-- 
---------------------------------------------------
Dr. Jochen Hub
Computational Molecular Biophysics Group
Institute for Microbiology and Genetics
Georg-August-University of Göttingen
Justus-von-Liebig-Weg 11, 37077 Göttingen, Germany.
Phone: +49-551-39-14189
http://cmb.bio.uni-goettingen.de/
---------------------------------------------------