Hi Charles,

Thanks for the information. I have used rdc.setEnsWeights and
rdc.useSimEnsWeight = False and the same for NOE.

Is there any test script or documentation available for the weight
optimization facility? We are working on a similar idea and would like to
know XPLOR's capabilities.

Thanks
Santhosh


On Wed, Sep 4, 2013 at 1:45 PM, Charles Schwieters
<char...@schwieters.org>wrote:

>
> Hello Santhos--
>
> >
> > This is more of confirmation than anything.
> > I know that RDCpot is a weighted average, where the weights provided by:
> >
> > #weights is a list of weights
> > esim = EnsembleSimulation("ensemble",ensembleSize)
> > memberIndex = esim.member().memberIndex()
> > esim.setWeights(weights)
> >
> > are used for computing the weighted average. Are NOE's too done this
> way? It
> > was not mentioned in the documentation.
> > Is there anything in particular that I should take care when using
> weights
> > apart from that one line of code esim.setWeights()?
> >
>
> The ensemble weights associated with the EnsembleSimulation are
> default weights for ensemble-aware potential terms with default values
> of 1/Ne. These weights are appropriate for AvePot-averaged terms,
> where each ensemble member is expected to have the same behavior. Each
> energy term also has a local ensWeights member, which can be set with
> setEnsWeights( sequence):
>
>   ensWeights      - a sequence of ensemble weights to use when calculating
>                     the ensemble- averaged value of each dipolar coupling.
> By
>                     default, the weights of the underlying
> EnsembleSimulation
>                     are used. If these are overridden by calling the
>                     setEnsWeights method, and useSimEnsWeight is set to
> False.
>
> This later mechanism is the correct way to tune ensemble weights, and
> is used in the weight-optimization facility.
>
> I hope this helps. The documentation leaves much to be desired, but I
> have updated the NOEPot and other pages for the next release.
>
> best regards--
> Charles
>
_______________________________________________
Xplor-nih mailing list
Xplor-nih@cake.cit.nih.gov
http://cake.cit.nih.gov/mailman/listinfo/xplor-nih

Reply via email to