I'd second Rasmus's point - unless you make the stereospecific assignment
based on direct data, then (although very handy and probably mostly
correct), purely NOE based stereospecific assignments offered by CYANA or
ARIA are not and should not be treated as gospel. It would be nice
therefore if the data model & analysis & format converter were clever
enough to allow one to link the swapped resonances in the imported
structures to to their original prochirally ambiguous resonance without
implying that the prochiral assignment was made. It would be even nicer
if CYANA talked data model. I see Wim's already on this route.
Brian
On Fri, 27 Jun 2008, Rasmus Fogh wrote:
> Dear Marie,
>
> You make perfect sense.
>
> A word of warning, though. CYANA assumes stereospecific assignment in all
> cases, and swaps to get the lowest target function values. If the target
> function values are similar, the choice will likely be pretty random.
> Personally I would not accept CYANA stereospecific assignments as absolute
> truth, but would make an individual decision in each case. There is too
> big a risk of introducing wrong information otherwise.
>
> Unfortunately I am not the right person to come up with a fix.
>
> 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 Thu, 26 Jun 2008, Marie Phelan wrote:
>
>> Thanks guys
>>
>> Unfortunately the way Ben suggested to swap the prochirals in analysis
>> doesn't work for me as bizarrely I'm having trouble even though my
>> prochirals AREN'T stereospecifically assigned within my analysis
>> project (!). Therefore swapping between Hga and Hgb is not available
>> in the resonance browser (only between Hg1 and Hg2).
>>
>> The reason I have this trouble regardless of stereospecific ambiguity
>> is down to cyana optimally swapping diastereotopic pairs which are not
>> already stereoassigned to achieve the lowest possible target function
>> value.
>>
>> It never occurred to me to change them back through CYANA but that did
>> work by checking the log and changing the .prot file in cyana to
>> account for the swaps listed there, for example;
>>
>> read test-final.prot
>>
>> atom swap "39 VAL QG1"
>> etc
>>
>> write test-swap.prot
>>
>> Importing the new .prot file in with the cyana assigned .peaks gives
>> me consistent ambiguous assignment throughout original data and that
>> coming from
>> cyana.
>>
>> However now I think about it, if cyana is determining the
>> stereospecific assignment of my diastereotopic pairs based on the
>> structures generated, why does format converter revert this back to
>> ambiguous 'a' and 'b' when importing this assignment back into my
>> project? surely I should be able to keep these assignments and my
>> original ambiguous 'a' and 'b' assignments can be stereospecifically
>> assigned to match. I can, now that I have consistent ambiguous
>> assignment between original and cyana data, go through and
>> stereospecifically assign them based on the cyana output but I'm
>> wondering whether it could be possible for format converter to do this
>> directly?
>>
>> (I hope that made sense)
>>
>> cheers
>>
>> Marie
>>
>>
>> Quoting Ben Goult <[log in to unmask]>:
>>
>>> Hi Marie
>>>
>>>> 1.) Before importing the spectra, use Format Converter M:Process:Clean up
>>> resonance application data, then select cyana and xeasy to remove all the
>>> application data.
>>>
>>> Whoops! I meant to say before "importing your shiftlist" not spectra. The
>>> shift list to import being your "-final.prot".
>>>
>>>
>>>
>>> Reading my first post back it wasnt very clear so here is a better(?)
>>> description. I would recommend as the quickest way to sort your problem that
>>> you delete the imported cyana shiftlist/peaklists and re-import.
>>>
>>> 1. In format converter go to "M:Process:Clean up resonance application data"
>>> and select "cyana". Repeat this again but this time select "xeasy". This
>>> step is the key.
>>>
>>> 2. In the directory where you did your Cyana run is a ".log" file, towards
>>> the bottom of this file there is a list of swapped atoms, like this;
>>>
>>> Atoms with consistent swapping in 20 or more structures:
>>> TF gap # 1 5 10 15 20
>>> 28 VAL QG1 QG2 0.3426 20 ******************** swapped
>>> 37 MET HG2 HG3 0.6050 20 ******************** swapped
>>>
>>> The ones marked "swapped" have been swapped (!) by cyana and it is these
>>> shifts that are going to cause problems when you re-import the lists back
>>> into Analysis. However, by swapping these in Analysis before the import it
>>> is possible to keep everything synchronised.
>>>
>>> 3. So in this case, where the 37 MET Gamma protons are swapped go into
>>> Analysis and using M:Assign:Resonances locate 37 Met. Then select HG2 and
>>> click "Swap Prochiral"; this will make the HG2 into HG3 (and swap the HG3 to
>>> HG2). For 27VAL you need to select HG1* and "swap prochiral" AND also CG1
>>> and "swap prochiral", this will stop them getting mixed up. You need to do
>>> this for all the swapped atoms in the list.
>>>
>>> 4. Then you can read in the shift list "xeasy-final.prot" and choose to
>>> "link resonances"
>>>
>>> 5. Finally, you can read in the peak list "xeasy-cycle7.peaks" and choose to
>>> "match resonances".
>>>
>>> N.B. The swapping the resonances before importing the lists only works if
>>> you do Step1, otherwise they will get swapped back.
>>>
>>> Hope that helps.
>>>
>>> Cheers
>>>
>>> Ben
>>>
>>>
>>
>>
>> --
>> The University of Edinburgh is a charitable body, registered in
>> Scotland, with registration number SC005336.
>>
>
Dr. Brian O. Smith ---------------------- B Smith at bio gla ac uk
Division of Biochemistry & Molecular Biology,
Institute Biomedical & Life Sciences,
Joseph Black Building, University of Glasgow, Glasgow G12 8QQ, UK.
Tel: 0141 330 5167/6459/3089 Fax: 0141 330 4600
|