positive Lennard-Jones energies
pelleque at SCRIPPS.EDU
Wed Aug 14 18:03:56 EST 1996
----- Begin Included Message -----
Fellow X-PLOR users--
I've been calculating structures from NMR data and have
discovered that all my structures have positive values for
the Lennard-Jones energy. This is surprising to me,
since all of the other statistics look relatively good--
rmsd's are low, all the other energy terms are reasonably
small, and no constraints or bond lengths or angles
are violated consistently.
The structures were calculated with distance geometry and refined
using the parallhdg.pro force field and the refine.inp protocol
(which uses a purely repulsive (REPEL) van der Waal's term).
I then transferred the structures to the PARAM19x and PARALLH22X
force fields and calculated the Lennard-Jones energies with the
TRUNCated potential and cutnb=50. Both force fields gave me
values greater than zero, ranging from 100-200 and 1000-3000
Has anyone else run into similar problems? If anyone has
suggestions or strategies for identifying the problem, I'd
----- End Included Message -----
The repel function is only repulsive (of course) and does not account
for the attraction of atoms. It has necessary higher energy than a
complete lennard-Jones (attraction term missing).
A simple way to see that is to minimize only side-chains of your
molecule without any experimental constraint (vdw will go down).
Nevertheless to fix your problem you may:
1) refine your model with a complete lennard-Jones potential and
2) Minimize slightly your current model using energy function as in 1).
More information about the X-plor