Hi Wayne,
This is exactly the right description. But are you saying that the
"circular" pick peaking is not currently implemented? If not, are there
any plans to add the option? It would be extremely useful for the folded
spectra and, I guess, not too difficult to implement the way you
described it. In NMRView there is also an option for an opposite sign
match across the edge to take into account a sign flip for certain
folding modes. In this case you need to consider x[1] > x[2] and x[1] >-
x[N].
Cheers
Igor
Wayne Boucher wrote:
> Hello,
>
> I think this is the edge effect, since my guess is currently that for row
> N it is only checking for maximality against row N-1 and not also against
> row 1 (which it should be, using circularity). (N = number of points in
> given dimension.) So we should be checking whether
>
> x[N] > x[N-1] and x[N] > x[1]
>
> (well, that's with the Fortran/Nmr convention where points go from 1 to N
> and not from 0 to N-1), and I think only the first is being checked.
> And similarly at the other edge, we should be checking whether
>
> x[1] > x[2] and x[1] > x[N]
>
> If both conditions are checked then you should only get one peak, not two,
> on the edge.
>
> Wayne
>
> On Fri, 16 Jun 2006, Ben Goult wrote:
>
>
>> Hi
>>
>> The tiling method is a bit problematic, as the current 4D spectra I am
>> looking at is 3 times folded in 2 of the dimensions, therefore when I use
>> the tiling I end up with 9x as many peaks which Analysis seems to struggle
>> to handle.
>>
>> So ideally I want to be able to try a different approach and only pick the
>> peaks in the folded spectrum (without unfolding it), and to do this
>> properly it would be nice if analysis would look for the real maxima of
>> peaks on the edge of the spectrum (which ever side it may be on). Is this
>> possible?
>>
>> I think this is a separate issue to the tiling error.
>>
>> Cheers
>>
>> Ben
>>
>>
--
Dr. Igor Barsukov
Biological NMR Centre,
University of Leicester
PO Box 138,
University Road,
Henry Wellcome Building,
Leicester LE1 9HN
UK
E-mail: [log in to unmask]
Tel: +44 (0)116 229 7098
FAX: +44 (0)116 229 7053
|