AMBER Archive (2008)

Subject: AMBER: Minimization in Octahedric Box

From: guardiani_at_fi.infn.it
Date: Tue May 06 2008 - 06:31:50 CDT

Dear Amber users,

I am trying to perform an MD simulation of a peptide in water in an
octahedric box using Amber 9. I prepared the topology with xleap
using the instructions:

x = sequence { NARG ASP HIE SER TYR GLN GLU CGLU }
solvateoct x TIP3PBOX 8.0
addions x Na+ 0
saveamberparm x mog_101-108_ext_watoct.top mog_101-108_ext_watoct.crd
savepdb x mog_101-108_ext_watoct.pdb
quit

I then tried to minimize the solvent while keeping the peptide restrained
in the initial conformation. I used the following input file:

Minimization
 &cntrl
  imin = 1,
  maxcyc = 5000,
  ncyc = 2500,
  ntb = 1,
  ntr = 1,
  cut = 12.0,
 /
Keep protein fixed
500.0
RES 1 8
END
END

Unfortunately at the end of the minimization run, I realized that the box
was deformed and in particular 6 satellite clouds of water appeared around
the main cloud surrounding the peptide (see attached files).
I then repeated the simulation several times trying to play with the
parameters. I tried all the following:

1) Removing the cutoff
2) Performing the simulation at constant pressure
3) Performing only 5000 steps of steepest descent
4) Performing only 500 steps of steepest descent followed by
   500 steps of conjugate gradient

The result, however, did not change.

In the hope to fix the geometry of the solvent box, I also performed
a constant volume equilibration:

 &cntrl
  imin = 0,
  ntx = 1, irest = 0,
  ntpr = 100, ntwr = 100,
  ntwx = 100,
  ntb = 1,
  nstlim =1000000, dt = 0.001,
  ntt = 3, gamma_ln = 1.0,
  ntr = 1,
  restraint_wt = 0.05,
  restraintmask=':1-8',
  nscm = 50,
 &end
 &wt type='TEMP0',istep1=0,istep2=20000,value1=10,value2=310.0,/
 &wt type='END'

After this stage I noticed that the the octahedric box had turned into a
strange shape more close to a parallelepiped rather than to an octahedric
geometry (see attached file).
My impression is that sander treats the octahedric box as if it were a
parallelepiped one.

Finally, I decided to repeat the simulation using a rectangular parallelepiped
solvent box:

x = sequence { NARG ASP HIE SER TYR GLN GLU CGLU }
solvatebox x TIP3PBOX 8.0
addions x Na+ 0
saveamberparm x mog_101-108_ext_watbox.top mog_101-108_ext_watbox.crd
savepdb x mog_101-108_ext_watbox.pdb
quit

The minimization run this time did not modify the parallelepiped shape of the
box and the result was completely satisfactory. I therefore conclude that
I must have made some mistake in the use of the solvateoct command but I
don't understand which one. Do you have any suggestion ? I need the octahedral
box because my final goal is to perform REMD simulations of a peptide in a
mixed water/hexafluoroacetone solvent (I am following the tutorial
"Setting up an MD simulation on DNA in mixed water-ethanol solvent"), so that
it is important to keep the number of solvent molecules to a minimum to save
computational time.

Thanks in advance for your help,

                                 Carlo Guardiani

----------------------------------------------------------------
This message was sent using IMP, the Internet Messaging Program.


-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber_at_scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo_at_scripps.edu