Hello,
Tim says:
I will probably make this easier in future versions, but for the moment
the trick is:
Sort the restraints by violation (e.g. Mean Viol).
Select the troublesome ones (click + <Shift> for the range) in the table.
Then [Show Peaks].
In the resultant peak selection table select all (<a>) then [Set Details]
and enter "VIOL run X" (or
whatever you like).
These peaks can easily be found in the NOE contributions table by sorting
on the details column.
:-)
For info, the REJECTED are the consistently violated restraints. This will
include artefacts, noise, minor conformations etc, but also sometimes
useful restraints which can be re-introduced when a structure is more
refined. - I have noticed especially that flexible termini (or whatever)
can sometimes give coincidentally matching chemical shifts, and provide a
relatively sensible conformation, pushing out the real stuff.
Wayne
On Sun, 13 Nov 2011, Krisztina Feher wrote:
> Dear All,
>
> I find the NOE Contributions window extremely useful, but I am missing one bit of information in it: marking of the peaks for which the corresponding restraints were violated during the calculation. It shows REJECTed peaks, but these are (correct me if I am wrong) just merged with their crossdiagonal pairs.
>
> The way I do it (suggest me a better strategy if you have one) I let ARIA assign the NOE peaks, let it reimport structures, peaks, restraints and rejected restraints. Then I look for peaks that caused violations or got rejected to reassign them or to set their Merit to zero, because they are artefacts. The NOE contributions are best for reassigning peaks, but it does not show peaks, for which the restraints were violated. So, the way around it is to open Restraints and Violations, select the reimported restraints from ARIA for which there is a violation list calculated against it8, click on the violated restraint, click on Show Peaks and look up that Peak number in the NOE Contributions.... This is just clumsy, I am sure there is a better way of doing it! Any one knows a better way of checking up peaks, which were causing problems during the calculations, I would appreciate any ideas.
>
> Thanks a lot,
> Krisztina
>
|