On Tue, 11 Apr 2006, Tim Stevens wrote:
>> in "Contraints":
>>
>> * Violations seem to be calculated per assignment-option rather than
>> per restraint (total r-6 average over all assignment-options)
>> which reports many contraints as violated that are not considered
>> as violated by CNS or Xplor-NIH.
>
> Not true. All ambigous items under the constraint are considered but only
> if they are ALL outside the range of a constraint will a violation be
> issued. (At least that's the intention)
>
> I could switch to calculating the violation value on the closest matching
> item (perhaps better anyhow), if that's what you mean. Otherwise a
> specific example would help.
I can see that there is a bit of a complex web of possibilities here, but
echoing Simon's thoughts, I would have thought that the most sensible
default is to use r-6 sum averaging over all possible distance pairs in a
restraint and report a violation only if the effective distance is longer
than upper limit. This is what most people are using in their structure
calculations and makes the most sense in almost all cases.
I can see a possible difference with how possibilities are displayed and
colour coded on the structure viewer (though I haven't checked the code to
be certain). At the moment I'm guessing that you colour code the various
inter-atom connections by reference to a distance, but perhaps something
that would be more in the spirit of r-6 sum averaging would be to colour
code by fractional "contribution" to the crosspeak volume/intensity based
on an ISPA (at least that would be a good starting model). Happy to
expand more if this isn't clear.
Brian
--
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 8640
|