[gmx-users] projection on the first two eigenvectors obtained from the concatenated trajectories

[Tsjerk Wassenaar]

Hi Qasim,

Didn't see the plot :) Now I did, and it's not strange. It's a dimer? I
think that the one/two ligand cases may be quite similar, but they look
different, because they're mirrored. You can imagine a line through the
center of wt and the joint center of the other two and see what I mean.
Also, the first and second eigenvectors as you got them are still mixed.
The difference between the wt/ligand case can be best described by that
line between the centers as I just sketched.

I found the tutorial of Lindsay Smith on PCA quite good. Google knows where
it is.

Cheers,

Tsjerk
On Jun 8, 2016 2:04 AM, "Qasim Pars" <qasimpars at gmail.com> wrote:

> Dear Tsjerk,
>
> Then my plot is not strange?
> You said that "The overall mean of any set of projections lies at the
> origin". Could you please suggest me any material that explains what you
> said? I couldn't find anything about that on google.
>
> Cheers,
>
> On 7 June 2016 at 15:15, Tsjerk Wassenaar <tsjerkw at gmail.com> wrote:
>
> > Hi Qasim,
> >
> > What makes you think that they should start at 0,0 ?
> > The overall mean of any set of projections lies at the origin.
> > PCA is a pretty powerful and advanced technique, that requires quite a
> bit
> > of thinking in the preparation, the execution and the interpretation.
> > Please do read plenty of tutorial material dealing with it, not only from
> > the perspective of simulations, but also from the linear algebra view.
> >
> > Cheers,
> >
> > Tsjerk
> >
> > On Tue, Jun 7, 2016 at 2:09 PM, Qasim Pars <qasimpars at gmail.com> wrote:
> >
> > > Dear gmx users,
> > >
> > > I have simulated three systems. The projection of each individual
> > > trajectory on the first two eigenvectors obtained from the concatenated
> > > trajectory is strange because each projection should start at (0,0)
> > points
> > > or near (0,0). Please have a look at the plot
> > > https://www.dropbox.com/s/3jrlkdwbv4csccq/pca.png?dl=0
> > >
> > > Why the projections do not start at (0,0) points or near (0,0)? Can
> > anyone
> > > of you tell me where I might have made a mistake?
> > >
> > > In order to remove jumps across the box of each trajectory I used the
> gmx
> > > trjconv -pbc nojump. The other commands are below:
> > >
> > > trjcat -f wt.xtc one-ligand.xtc two-ligands.xtc -o merged.xtc -cat
> > >
> > > gmx covar -f merged.xtc -s wt.pdb
> > >
> > > gmx anaeig -f wt.xtc -s wt.pdb -v eigenvec.trr -2d wt.xvg -first 1
> -last
> > 2
> > > gmx anaeig -f one-ligand.xtc -s wt.pdb -v eigenvec.trr -2d
> one-ligand.xvg
> > > -first 1 -last 2
> > > gmx anaeig -f two-ligands.xtc -s wt.pdb -v eigenvec.trr -2d
> > two-ligands.xvg
> > > -first 1 -last 2
> > >
> > > Cheers,
> > >
> > > --
> > > Qasim Pars
> > > --
> > > Gromacs Users mailing list
> > >
> > > * Please search the archive at
> > > http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before
> > > posting!
> > >
> > > * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
> > >
> > > * For (un)subscribe requests visit
> > > https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or
> > > send a mail to gmx-users-request at gromacs.org.
> > >
> >
> >
> >
> > --
> > Tsjerk A. Wassenaar, Ph.D.
> > --
> > Gromacs Users mailing list
> >
> > * Please search the archive at
> > http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before
> > posting!
> >
> > * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
> >
> > * For (un)subscribe requests visit
> > https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or
> > send a mail to gmx-users-request at gromacs.org.
> >
>
>
>
> --
> Qasim Pars
> --
> Gromacs Users mailing list
>
> * Please search the archive at
> http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before
> posting!
>
> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>
> * For (un)subscribe requests visit
> https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or
> send a mail to gmx-users-request at gromacs.org.
>