Dear Karthik,
I think it is a bit dangerous to change the data in the way you propose -
Wims method and 0.001 ppm cutoff sounds safer.
It is true that these small differences probably come from peak averaging,
but if it is never possible to distinguish, all peaks ought to be assigned
to both possibilities. Except for errors of course.
In practice you are probably right - a difference of 0.001 ppm is unlikely
to be reliable. On the other hand you are not doing any damage if you give
shifts that differ by 0.001 when the uncertainty is 0.01 or higher anyway.
Rather than fixing an arbitraty lower cutoff for assignment differences,
we are better off keeping the data as they were entered. We could add a
test for this situation in the assignment validation routine (right,
Tim?), and let people check for themselves what they want to report.
Yours,
Rasmus
---------------------------------------------------------------------------
Dr. Rasmus H. Fogh Email: [log in to unmask]
Dept. of Biochemistry, University of Cambridge,
80 Tennis Court Road, Cambridge CB2 1GA, UK. FAX (01223)766002
On Wed, 2 Feb 2011, Karthik Rajasekar wrote:
> Hi Wim,
>
> I don't think I have a resoultion in my sepctra for 0.001 ppm. I think this
> should be fairly a general case. I think I have assignments as 2.019 and
> 2.020 probably beacuse of averaging over multiple spectra and in reality they
> are the same peak. I think 0.01 would be a good general cutoff.
>
> Would be nice to see what others think too.
>
> Karthik
>
>
> Wim Vranken wrote:
>> Hi all,
>>
>> To follow up on this, the chemical shifts had to be exactly the same (to
>> the n-th decimal) for the ambiguity code to be 1. I've now changed this so
>> that it has an accuracy of 3rd decimal place, which means that this:
>>
>>> 42 5 Leu HB2 H 1.525 0.004 2
>>> 43 5 Leu HB3 H 1.525 0.004 2
>>
>> will become:
>>
>>> 42 5 Leu HB2 H 1.525 0.004 1
>>> 43 5 Leu HB3 H 1.525 0.004 1
>>
>> but this:
>>
>>> 3033 246 Glu HB2 H 2.019 0.000 2
>>> 3034 246 Glu HB3 H 2.020 0.003 2
>>
>> will stay the way it is.
>>
>> Hope that works for most,
>>
>> Wim
>>
>> PS This is not yet on the update server, but will be soon!
>
>
> --
> ##############################################################
> Dr. Karthik Rajasekar Ph: 01223766015
> Research Associate M : 07807212010
> Department of Biochemistry Fax: 01223766002
> University of Cambridge
> Cambridge CB2 1GA
> ##############################################################
>
|