[gmx-users] Re: different trajectories

[David van der Spoel]

On Tue, 26 Feb 2002, Bruno Rizzuti wrote:

>the hereafter reported answer you gave to a Gromacs user arose me a
>question. When you speak about "different trajectories" I guess you mean
>not *completely* different
 otherwise, it could mistrust the whole
>meaning of simulations. So my question is: what kind of parameters do
>you look at, to see whether two different simulated protein structures
>or two different trajectories are sufficiently "similar" not to be
>worried about the inevitable little differences (for instance
>differences coming from altering the seed number, but also coming from
>using different cut-off radii, number of solvent molecules, etc.)? RMSD
>and RMSF of protein atomic position are examples of suited parameters?
>And what about others?
For liquid simulations it is easy, you look at energy, density and things
like diffusion constants, dielectrics etc. For *sufficiently long* 
simulations these numbers should be identical and this then proves that 
your simulation is in equilibrium. 

For proteins it is much more difficult, because you only have one molecule 
in solution. Here one could check indeed the RMSF, RMSD from starting 
structure, secondary structure etc. However since you only have one 
molecule there will be most likely be differences between the simulations. 
In principle an infinitely long simulation of a single molecule will give 
you the same information as a shorter one with many molecules, but 
infinite is long...

>By the way, the number 173529 is not a prime number, for instance it can
>be divided several times by 3
 so I wonder why that number is present in
>the examples.
Lazyness, and it is not that important... 

Groeten, David.
________________________________________________________________________
Dr. David van der Spoel, 	Biomedical center, Dept. of Biochemistry
Husargatan 3, Box 576,  	75123 Uppsala, Sweden
phone:	46 18 471 4205		fax: 46 18 511 755
spoel at xray.bmc.uu.se	spoel at gromacs.org   http://zorn.bmc.uu.se/~spoel
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++