I should have added one more thing. The system under investigation was a
complex of a protein and a general organic ligand, parameterized using
GAFF. The ambertools handbook says
"Force field calculations (e.g. molecular dynamics and minimization) can
be carried out with an implementation of the AMBER force field."
I assumed, this means all amber force-fields, including GAFF and GLYCAM.
Maybe this causes the "nan".
Jochen
David A. Case schrieb:
> On Mon, Feb 09, 2009, Jochen Heil wrote:
>
>
>> For some reason, i get "nan" for the
>> total energy as well as the nonpolar part. Since i switched off GBSA, i
>> would expect zero, but "nan" seems to point to a wrong setup or logical
>> error, especially when printed in the total energy column.
>>
>
> I haven't ever seen anything like this. The nan for "total" just arises
> because it is the sum of the other components. And, since MD depends
> only on the forces (not the total energy), and the forces look OK, it's
> possible that the trajectory is OK as well. But this should be tracked
> down...it looks like a bad initialization or memory overwrite(?).
>
> What happens if you add a call to mme() after a (short) MD run? You
> might want to try a debugger or post a small example that allows the
> problem to be reproduced.
>
> ....dac
>
>
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
>
>
--
Dipl.-Ing. Jochen Heil
AG Kast
Eduard-Zintl-Institut für Anorganische und Physikalische Chemie
Technische Universität Darmstadt
Petersenstr. 20, 64287 Darmstadt, Germany
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Wed Feb 11 2009 - 01:17:31 PST
