[gmx-users] Re: mdrun -rerun and box size bug not fixed?

[Michel Cuendet]

Hi,

Thanks David, it helps when you know there is light at the end of the 
tunnel.

Using PME didn't have any effect, but using a full precision trajectory 
reproduced energies up to 5e-3 kJ/mol. So the imprecision of the xtc 
coordinates causes the inaccuracy of the energies, i.e. creates a 
systematic bias towards high energies. This can actually be explained by 
the fact that bonded potential energy functions have a positive second 
derivative, as well as the Lennard Jones potential in the region which 
contributes most to the total energy. It is easy to see that a random 
and symmetric noise epsilon on the position x (truncation error in the 
xtc file) gives values of U(x+epsilon) whose average is larger than U(x) 
if U''(x) > 0.

As far as I am concerned, I don't have the trr files anymore for my 
system. I can forget about accurate energies... But it's good to know 
that the effect exists, and can lead to biases of the order 2.5% on 
bonded energies with the xtc precision.

Bye,
Michel

>So I guess what I'm saying is that in principle you should be able to
>get good energies with rerun. Maybe you should try and narrow down the
>problem by doing some testing:
>(a) Do some reprocessing with PME rather than RF to see if that helps.
>(b) Do some reprocessing on full precision trajectories rather than
>xtcs to rule out precision issues
>etc.
>
>Best wishes,
>David
>
==========================================================
Michel Cuendet, Ph.D
Molecular Modeling Group
Swiss Institute of Bioinformatics
CH-1015 Lausanne, Switzerland
www.isb-sib.ch/groups/Molecular_Modeling.htm
==========================================================