Truncated octahedron - unwrapping the trajectory etc

From: Alexander Predeus (
Date: Tue Aug 16 2011 - 21:46:22 CDT

Hello all,

I've run into some problems trying to unwrap the trajectory that I've
obtained using NAMD. I've used the following (admittedly non-standard)
cell basis vectors for truncated octahedron-shaped cell:


where a=d/3*sqrt(3), and d is unit cell size (distance between
hexagonal faces in case of truncated octahedron).

While these are different from what's specified in GROMACS manual and
other sources, the angles between these vectors are all 109.47122 and
their length is d, so they are right vectors.

The reason I chose those is to match the setup of "crystal" facility
in CHARMM - it's very convenient to analyze trajectories obtained by

Now, to questions.

1) are there any other ways of unwrapping the trajectories in NAMD
except PBCtools in VMD? It's a CHARMM format of trajectory I'm talking
about. I have trouble installing vmd-1.9 and vmd-1.8.7 has a described
bug in pbctools preventing me from using cell basis vectors not
aligned with x axis.
2) after I've ran the simulations, I've checked the periodic cell
information written in dcd file's frames. While it should be something

76.26092 x 76.26092 x 76.26092 ( 109.47122 x 109.47122 x 109.47122 )
(obtained from CHARMM simulation)
NAMD writes something of this kind:
76.38898 x 76.38898 x 76.38898 ( 90.49894 x 90.49894 x 90.49894 )
the angles change slightly, but they stay around 90 degrees. This
seems wrong; also, this prevents the correct unwrapping using CHARMM

Is this a bug, or does it have some explanations?

3) to run the simulations, I used

wrapWater on
wrapAll on
wrapNearest on

if in the future I want to prevent segments of my protein from being
separated, do I just turn "wrapAll" off? Do I have to also turn
"wrapWater" off?

Thank you.

-- Alex Predeus
Michigan State University

This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:57:36 CST