[gmx-users] Molecular geometry depends on the orientarion

[T. Zaraiskaya]

  Hello,

 I am simulating Galactosyl Ceramide lipid bilayer using
 ffG43a2x ff. I've built the topology and as a first test
 I am simulating the crystal of this lipid.

 Surprisingly, when I performed energy minimization it turned out
 that the optimized molecular geometry depends on the orientation of the
 molecule.
 I have attached the picture, showing experimental X-ray
 structure (yellow) and optimized structure (red).

 As you can see, for the upper molecule deviation between the
 experimental and minimized coordinates is rather small and  sugar
 impropers are reproduced correctly. However, the lower part of the
 crystal looks as if  all sugar impropers would have changed their sign.
 Note that the upper yellow molecule is the SAME molecule as
 the lower, it is just rotated 180 around x-axis.
 Why the results of minimization are different? I can make
 mdrun to predict the correct conformation of the lower part of the
 crystal by changing signs of the sugar impropers, but then of course the
 upper part does not look good. Is it a bug in the program or I am doing
 something silly?


  Tanya


-- 
====================================
Tanya Zaraiskaya
Physics Department
University of Guelph
Guelph, ON, N1G 2W1
CANADA
email:tz at helios.physics.uoguelph.ca
phone:(519)824-4120 x8346.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: plot2.jpg
Type: image/jpeg
Size: 24179 bytes
Desc: 
URL: <http://maillist.sys.kth.se/pipermail/gromacs.org_gmx-users/attachments/20020710/794e7d68/attachment.jpg>