On Fri, 10 Nov 2006, eiso wrote:
>> Wouldn't the geometric mean be more appropriate/informative for this type
>> of situation?
>>
>
> Brian, could you a specific example of why you think this is the case?
> I don't understand
Not particularly - I'm with you that we need the r-1/6 sum most urgently
for the purpose of comparison with NOE data and that this is going to be
the most useful value for most analysis users. I can sort of imagine that
there is a case in there, as Rasmus says, for wanting to know something
that relates to the real distance between groups of atoms in a structure
model, though for most of these cases you'd probably be looking at your
structure in <insert favourite molecular graphics package>. For those
cases (i.e. not comparing with NOE data), you don't want a weighted
average/sum of the distances and the minimum distance also has
deficiencies. The geometric average (or distance between pseudo-atoms -
should come out the same I think) would be what you wanted.
We want r-1/6 sum as default 'cos we mostly compare model distances to NOE
data in analysis for the purposes of violation analysis/iterative manual
assignment. The option of minimum distance might also be useful for some.
r-1/6 averaging should be buried in a deeply inaccessible place if it's
left in at all.
--
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
|