From: Urahata (surahata_at_gmail.com)
Date: Thu Jul 06 2006 - 14:39:22 CDT

Hi all,

I appreciate if anyone could shed some light or point out to the right direction
for three basic operational questions regarding VMD.

1. We have a box with one protein solvated by explicit water. We want to
   show water molecules that are within a distance range from one residue.
   It is not just the closest waters but only the ones located between distances
   r and r+dr. Let's say that the range is defined by 5A and 8A. I tried :

   water within 8A of resname PHE and not water within 5A of resname PHE

   but that's not the right way.

2. Is it possible to load two (close) atoms without having VMD
(automatically) building
   a bond between them ?

   This question is related to how to build local density maps surround
   a reference site. We have a generic PDB file like this :

MODEL 1
ATOM 1 N REF 1 -13.599 2.432 13.582 0.00 0.00
ATOM 2 O HET 2 -13.273 2.670 11.413 0.00 0.00
ATOM 3 O HET 3 -12.623 2.466 12.636 0.00 0.00
ATOM 4 O HET 4 -14.717 2.852 12.940 0.00 0.00
ATOM 5 O HET 5 -14.397 3.426 11.736 0.00 0.00
...
ATOM 1000 H HET 1000 -13.397 1.426 10.736 0.00 0.00

   where the first atom is the reference site and all other sites (O)
are positions that occurred during all simulation (ie, this is the equivalent
of loading all snapshots of closer neighbors at once). Right now, VMD loads
a bunch of them (although it complains about the maximum number
of bonds) until it reaches a limit of complaints and them halts the operation.
   I believe that one alternative is to repeat the reference site for
each neighbor and them load them all at once, as in a "movie",
but I would like to avoid redundancies and waste of disk space.

3. Is it possible to load, let's say, water density in some sort of grid
   for later renderization ? This would lead to the same goal of question 2,
   without loading too many overlapping points/sites.

   Thank you all,

   Sergio