[gmx-users] R.M.S.I.P for identical eigenvectors not equal to 1 ?

[Tsjerk Wassenaar]

Hi Raja,

I really don't get it. From the definition it follows directly that if
you have the same sets of eigenvectors, the RMSIP will be 1. Each
eigenvector from a set gives an inproduct of 1 with itself and of 0
with any of the others, yielding a total of 10 for the double sum...
etc. If you use the same set twice, you probably have a bug in your
script.

Besides, from the definition it also follows that the RMSIP can not
exceed 1. I think 3.5% error is quite a bit.

Tsjerk

On 7/28/06, raja <raja_28 at fastmail.us> wrote:
> Hi Tsjerk,
> Thanks for your reply. I used identical values of 10 eignevectors in two
> different file to compute r.m.s.i.p. I have pasted the link of the
> article from where I used the formula for r.m.s.i.p. More for some other
> cases of non-identical values, it slightly exceeds the value of 1 (e.g.,
> 1.0352), Which I am thinking due to some numerical problem. But I am
> confident that my script exactly following what the formula indicated in
> the article as pasted hereunder.
> http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=1304087
>
> With thanks !
> B.Nataraj
>
> On Fri, 28 Jul 2006 11:32:04 +0200, "Tsjerk Wassenaar"
> <tsjerkw at gmail.com> said:
> > Hi Raja,
> >
> > Did you use the same eigenvectors from the same simulation over the
> > same stretch of time? For two (parts of a) simulation(s) of the same
> > system, the RMSIP value will only converge to one in the limit of
> > infinite time, or on long time scales at least. The eigenvectors
> > obtained from two different (parts of) simulations can vary if
> > different parts of the conformational space are sampled or simply if
> > the sampling is different, which will be almost always for a limited
> > stretch of time.
> >
> > Best regards,
> >
> > Tsjerk
> >
> > On 7/28/06, raja <raja_28 at fastmail.us> wrote:
> > > Dear all,
> > >
> > > I have written small perl script to compute R.M.S.I.P.(Root mean square
> > > internal product) for comparing two eigen vectors set from two different
> > > simulation. I compared first 10 eigen vectors of each set. The strange
> > > thing is that while I compare eigen vectors of same simulation, it is
> > > showing a value of 0.54, but literature says, r.m.s.i.p of 1 indicates
> > > that the sets are identical, while a value of 0 indicates that the
> > > eigenvectors are orthogonal. I also confirmed my script validity by
> > > manually computing the R.M.I.P for small set of values.
> > >
> > > Could any body explain me a bit regarding this issue?
> > >
> > > With thanks!
> > > B.Nataraj
> > > --
> > >   raja
> > >   raja_28 at fastmail.us
> > >
> > > --
> > > http://www.fastmail.fm - A fast, anti-spam email service.
> > >
> > > _______________________________________________
> > > gmx-users mailing list    gmx-users at gromacs.org
> > > http://www.gromacs.org/mailman/listinfo/gmx-users
> > > 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/mailing_lists/users.php
> > >
> >
> >
> > --
> >
> > Tsjerk A. Wassenaar, M.Sc.
> > Groningen Biomolecular Sciences and Biotechnology Institute (GBB)
> > Dept. of Biophysical Chemistry
> > University of Groningen
> > Nijenborgh 4
> > 9747AG Groningen, The Netherlands
> > +31 50 363 4336
> > _______________________________________________
> > gmx-users mailing list    gmx-users at gromacs.org
> > http://www.gromacs.org/mailman/listinfo/gmx-users
> > 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/mailing_lists/users.php
> --
>   raja
>   raja_28 at fastmail.us
>
> --
> http://www.fastmail.fm - Does exactly what it says on the tin
>
> _______________________________________________
> gmx-users mailing list    gmx-users at gromacs.org
> http://www.gromacs.org/mailman/listinfo/gmx-users
> 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/mailing_lists/users.php
>


-- 

Tsjerk A. Wassenaar, M.Sc.
Groningen Biomolecular Sciences and Biotechnology Institute (GBB)
Dept. of Biophysical Chemistry
University of Groningen
Nijenborgh 4
9747AG Groningen, The Netherlands
+31 50 363 4336