[gmx-users] Blowing up a 512 dodecanol system ..

[Sergio Garay]

Dear Gromacs Users

I've prepared a 1-dodecanol system based on an 1-octanol-GAFF topology
taken from:
http://virtualchemistry.org/molecule.php?filename=1-octanol.zmat

I've run a minization of one dodecanol and also a box with 512 molecules
without any visible problem. I did not find any weird distance, angle o
dihedral angle.

The problem starts when I try to run a NVT simulation because the system
blows up. I have watched the time evolution (VMD) of the system every 2 fs
(because it blows up at the very beginning), I could see that most of the
molecules have a proper conformation, but one of them suffers a dramatic
(no sense) conformational change. For example, two hydrogens in one
methylene group became very close and even they overlap each other. After
one or two steps from this event the system blows up.

I've tried with no success:
* I have read the gromacs web page where it is described all the details to
take into account when a Blowing up happen (
http://www.gromacs.org/Documentation/Terminology/Blowing_Up)
* Check the topology file, looking carefully for some missing term
* Decrease the temperature of the simulation.
* Increase the temperature coupling time
* Decrease the integration time step

I've run several times this system and the molecule which blows up is
always different, but the rest of the molecules look healthy (no weird
conformation).

I need an equilibrated box of dodecanol (compatible with Amber forcefield)
in order to run a simulations of one the molecules that I am interested in.

Any help will be very appreciated.


-- 








*Sergio GarayDr. en Ciencias BiológicasFacultad de Bioquimica y Cs.
BiológicasUniversidad Nacional del LitoralSanta Fe - ArgentinaC.C. 242 -
Ciudad Universitaria - C.P. S3000ZAAArgentinaPh. +54 (342) 4575-213Fax. +54
(342) 4575-221 *