Dear Tharin,
That does make sense, but there is a slight information loss. (Always
assuming that the distances are similar. If they differ by more than 1.8A
for a CH2 group, of course, you would be better off taking the shorter
one, and setting it to either proton.)
Ideally you should, as part of the protocol, swop the assignment around
randomly during structure generation, so that Hba maps to HB2 in some
cases and to HB3 in others. Some programs instead swap the actual
chirality of the prochiral groups, which achieves the same effect (but can
be tricky to clean up). Either way you end up having to do statistics over
an ensemble where the same restraint is mapped to different atoms in
different structures. The best way to do it is a question of the structure
generation programs, not CCPN. Meanwhile your procedure is both simple and
sensible.
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 Tue, 30 Jul 2013, Tharin Blumenschein (CHE) wrote:
> Hi Rasmus,
>
> I think that in one of the examples, Sylvain had NOEs to *both* protons,
> looking at the restraints that he emailed originally.
>
> For 20SerHb2 and Hb3, where the chemical shift for both protons are the
> same, the restraint should be to either - we have no way of really knowing
> if one or both are close enough to give an NOE.
>
> But in the case of 13AspHbb and Hba, there are restraints for both, with
> different distances, suggesting there are actually NOE crosspeaks for both
> HB atoms. In this case, I'd want to keep both restraints. What I have done
> in the past is to set restraints to both atoms, but with the larger
> distance value of the two, given that we don't know which one has the
> closest distance. Does that make sense?
>
> Best wishes,
>
> Tharin
> -----------------------------------------
> Dr. Tharin Blumenschein, Lecturer
> School of Chemistry
> University of East Anglia
> Norwich - NR4 7TJ
> United Kingdom
> Phone: +44 (0)1603 59 2963
>
>
>
>
>
> On 29/07/2013 11:01, "Rasmus Fogh" <[log in to unmask]> wrote:
>
>> Dear Sylvain,
>>
>> As it happens, I disagree with you about the ambiguous case. What you
>> know
>> for sure is that the NH (for example) is close to *either* of the protons
>> - that is enough to give a signal. This limits the NH position to one of
>> two spheres centered on either proton. It is true that what is close to
>> one is likely close to the other as well, but that should follow
>> naturally
>> from covalent geometry. The volume that is close to *both* protons is
>> clearly more restricted than the volume that is close to either of them -
>> it is the intersection of the spheres, as opposed to the union. If you
>> expand to nuclei with a wider separation, the problem is more obvious.
>> Tyr/Phe aromatic protons or Leu/Val methyl groups are separated by over
>> 2A. Forcing your NH (or whatever) to be close to *both* could move it by
>> over 2A.
>>
>> 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 Mon, 29 Jul 2013, Sylvain Broussy wrote:
>>
>>> Dear Rasmus,
>>>
>>> Thank you very much for your answer.
>>> You are probably right for the stereospecific/non-stereospecific case:
>>> XPLOR may have a way to handle it by swapping the two atoms. It is more
>>> a question about XPLOR than the Format Converter itself, but could
>>> someone please confirm the fact, and tell us the method name and if it
>>> is included by default in the last version available for download?
>>>
>>> However, for the ambiguous case, I think the program should convert to
>>> *both* HB2 and HB3 (and not "OR"): the two atoms have very similar
>>> environments (they have very similar chemical shifts), and it is
>>> difficult to believe that one of them is close to the NH when the other
>>> one is not? This is how I understand the description of the ambiguous
>>> case in the "atom browser" section of the online documentation at:
>>>
>>> http://www2.ccpn.ac.uk/documentation/analysis/popups/BrowseAtomsPopup.htm
>>> l
>>> but I may have missed something?
>>>
>>> Sylvain
>>>
>>>
>>>> Dear Sylvain,
>>>
>>> Wim Vranken will have to give the final answer, but meanwhile I can
>>> maybe
>>> help a bit:
>>>
>>> For the ambiguous case (HB2 and HB3 have the same shift) I think the
>>> treatment is correct. If you see a peak, that generally means that your
>>> HN
>>> is close to *one* of the HB, but you do not know which (how could you -
>>> they have the same shift). Converting to HB* would seem to be OK. What
>>> do
>>> you think the program should do here?
>>>
>>> For the stereospecific and non-stereospecific case there is the problem
>>> that XPLOR does not have different atom names for two cases, as far as I
>>> know. Some programs use stereospecific atom names and then swops the
>>> chirality around during structure determination. I admit that I do not
>>> know how XPLOR distinguishes between stereospecific and
>>> non-stereospecific, but for what it is worth I suspect that you would
>>> have
>>> atom names HB2 and HB3 in eitehr case.
>>>
>>> That is all from me. Over to Wim,
>>>
>>> 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 Mon, 15 Jul 2013, Sylvain wrote:
>>>
>
|