[gmx-users] regarding converge of system

[Tsjerk Wassenaar]

Hey :)

If I say something is commonly used, I don't mean that the usage is correct
:p Chris' answer is quite good though, and I can't add much more, except
maybe that the RMSD, being a distance, isn't a good indicator of
convergence, when it reaches the point of saying 'far away'. There's just
too much far away; the conformational space associated with an RMSD of 0.7
or more is huge! So if a structure stays at that distance, it may still be
flying around like there's no tomorrow. The convergence of the RMSD against
the average structure will be better, and so is the drift in the average
structure itself. Note though, that these may still show apparent (or
false) convergence if they stay in a particular region of conformational
space.

Cheers,

Tsjerk



On Thu, Nov 1, 2012 at 3:13 PM, Christopher Neale <
chris.neale at mail.utoronto.ca> wrote:

> RMSD of what? Probably you mean RMSD from the starting (or crystal)
> structure. First, consider that your profile
> of RMSD vs. crystal structure levels off at 0.4 nm with increasing
> simulation time. Consider how many possible
> conformations are 0.4 nm RMSD away from the crystal structure. A stable
> RMSD does not necessarily indicate
> that you have obtained Boltzmann sampling of a particular conformational
> basin. Very generally, it is the
> degeneracy of RMSD with respect to conformational space that makes RMSD an
> insufficient indicator of
> convergence. There are other ways to look at this though. For instance,
> there might be an important degree of
> freedom that is important for your system but does not greatly affect the
> global conformation (and thus the
> RMSD). This might be a dihedral angle, formation of a hydrogen bond,
> tilt/rotation of a protein/peptide in a
> lipid bilayer, solvation patterns, ligand binding, system volume, etc.
>
> Also, and I presume that this is not what Tsjerk meant (but you should
> check whatever reference you are referring
> to) it is always possible that you remain stuck within a conformational
> basin that is locally favourable but globally
> unfavourable.
>
> For other measures of convergence, first think about what you are trying
> to study for your system and use that
> to guide your selection of properties for which you can evaluate
> convergence. This should include both global
> and local structural measures. There are loads of papers available that
> discuss this. See papers by Grossfield,
> Zuckerman, and others. A colleague of mine has also shown that some
> properties converge faster than others
> (meaning that some properties, like the RMSD, converge before the
> partition function has converged:
> http://pubs.acs.org/doi/abs/10.1021/ct900302n ).
>
> Chris.
>
> -- original message --
>
> I have two questions:
>
> 1-We can say "The RMSD is commonly used as an indicator of convergence
> of the structure towards an equilibrium state (Tsjerk W.)". RMSD is
> not sufficient to determine whether or not converge of
> structure/system. Why?
>
> 2-Radius of gyration, RMSD, rmsd matrix are used as an indicator of
> convergence of the structure. are there other indicators of
> convergence of the structure?
>
> Thanks in advance
>
> --
> Ahmet Yıldırım
> --
> gmx-users mailing list    gmx-users at gromacs.org
> http://lists.gromacs.org/mailman/listinfo/gmx-users
> * Please search the archive at
> http://www.gromacs.org/Support/Mailing_Lists/Search before posting!
> * Please don't post (un)subscribe requests to the list. Use the
> www interface or send it to gmx-users-request at gromacs.org.
> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>



-- 
Tsjerk A. Wassenaar, Ph.D.

post-doctoral researcher
Biocomputing Group
Department of Biological Sciences
2500 University Drive NW
Calgary, AB T2N 1N4
Canada