Re: initial estimate of the diffusion tensor -- December 12, 2007 - 21:15

It can be done, but I don't like the approach very much.  It is a way
of roughly mimicking the approach used in Tensor.  Nevertheless, see
the section:

6.4.1 Cytochrome c2 – optimisation of the diffusion tensor

in my PhD thesis on page 209.  I have no scripts to do this, but the
protocol is very basic.  This section and the corresponding methods
section describe fully how to implement this.  If you don't have the
link handy, the thesis is at
http://eprints.infodiv.unimelb.edu.au/archive/00002799/.

Regards,

Edward




On Dec 12, 2007 8:58 PM, Douglas Kojetin <douglas.kojetin@xxxxxxxxx> wrote:
> Hi Edward,
>
> Thanks for the detailed information.  The reason i asked about the
> estimation of the diffusion tensor was for the analysis of 3
> relaxation data sets only (r1, r2 and noe at a single field).  Is it
> possible to make an estimate for the diffusion tensor using relax?
>
> Thanks,
>
> Doug
>
>
>
> On Dec 12, 2007, at 2:43 PM, Edward d'Auvergne wrote:
>
> > Hi,
> >
> > The article you reference is a description of the use of the
> > 'full_analysis.py' script distributed with relax.  This script
> > implements a new model-free protocol which is described in Chapter 6
> > of my PhD thesis
> > (http://eprints.infodiv.unimelb.edu.au/archive/00002799/) and which
> > should be published very, very soon.  You can use the diffusion
> > tensor(s) from the Tensor program in relax but be careful.  As the
> > convention for the spherical and Euler angles used in Tensor are
> > undocumented I was unable to determine, even after asking, what they
> > are.  In Tensor I don't know how the symmetries are handled and
> > collapsed, whether 0 degrees for the azimuthal angle and alpha and
> > gamma is along x or y, whether the polar angle and beta range between
> > -pi/2 to pi/2 or between 0 and pi, or whether the Euler angles are in
> > the standard physics z-y-z notation, in the shifted axis system
> > z-y'-z', z-x-z, or roll-pitch-yaw, etc.  These are all nicely
> > described on Wikipedia.
> >
> > The point of the new protocol is that the model-free parameters are
> > determined before the global diffusion tensor.  This is in contrast to
> > the current procedure where you start with the diffusion tensor then
> > finally find the model-free parameters.  Hence the new protocol does
> > not use the initial tensor estimate from Tensor (this is an estimate
> > as many assumptions about the internal motions are made and mobile
> > residues should be filtered out first).  Note though that this
> > protocol minimally requires 4 relaxation data sets and optimally 6,
> > hence to use it you cannot avoid collecting at multiple field
> > strengths.  I hope this answers you questions.
> >
> > Regards,
> >
> > Edward
> >
> >
> >
> > On Dec 12, 2007 12:51 PM, Douglas Kojetin
> > <douglas.kojetin@xxxxxxxxx> wrote:
> > > Hi,
> > >
> > > In the following post [https://mail.gna.org/public/relax-users/
> > > 2007-05/msg00009.html], Edward said that relax could be used to
> > > estimate the diffusion tensor parameters (instead of using the
> > > program Tensor).  Does this protocol already exist in the form of a
> > > python script?  If not, could someone briefly outline the protocol in
> > > an email so I can write it myself?
> > >
> > > Thanks,
> > > Doug
> > >
> > > _______________________________________________
> > > relax (http://nmr-relax.com)
> > >
> > > This is the relax-users mailing list
> > > relax-users@xxxxxxx
> > >
> > > To unsubscribe from this list, get a password
> > > reminder, or change your subscription options,
> > > visit the list information page at
> > > https://mail.gna.org/listinfo/relax-users