Yes, in principle, on paper it is possible. Moreover in many cases by looking 
at the various directional Wilson plots you may be able to see direction of 
helices (just like in DNA/RNA). However in general case it is a little bit 
tricky (mixture of different secondary structures directed in different 
directions, noisy data etc). And there is another complication that bonds are 
involved in intensity curves via sinc function that is not completely local. I 
think there were attempts (by Hamburg group) with some success. What happened 
to their publication, I do not know.
You are right that in molecular replacement they are implicit. However if bonds 
are directed more or less in the same direction then assumption behind Wilson 
distribution breaks down and it may affect molecular replacement (not as severe 
as pseudo translation, nevertheless serious enough). Fortunately most molecular 
replacement programs are forgiving for such departures from assumptions.


regards
Garib



On 9 May 2012, at 20:28, Jacob Keller wrote:

> It seems to me that spherical forms of Wilson plots could be used to 
> determine how many bonds of what nature were oriented in which direction, and 
> this may have been what Bricogne's micro molecular replacement technique was 
> capitalizing on? For example, one might be able to orient a straight DNA 
> molecule by finding the direction at which the ~3.2-3.5 Ang bin signal was 
> greatest. But I guess this is probably implicit in molecular replacement 
> anyway...
> 
> JPK
> 
> On Wed, May 9, 2012 at 2:08 PM, Nat Echols <nathaniel.ech...@gmail.com> wrote:
> On Wed, May 9, 2012 at 11:58 AM, Garib N Murshudov
> <ga...@mrc-lmb.cam.ac.uk> wrote:
> > As far as I know there are several bumps: around 3.5-4 (there are some at
> > low resolution related with molecular shapes also) - secondary structures,
> > ~2.2 related with angles and around 1.2 related with covalent bonds. For
> > DNA/RNA there is one more bump around 1.6-1.7 ( I thought that is because of
> > Phosphor bonds). They are visible with normalised data better.
> 
> It has been pointed out to me that my example cut off the data at too
> low a resolution to see the peak for covalent bonds.  Here is a
> different version that shows a distinctive peak around 1.15A.
> 
> -Nat
> 
> 
> 
> -- 
> *******************************************
> Jacob Pearson Keller
> Northwestern University
> Medical Scientist Training Program
> email: j-kell...@northwestern.edu
> *******************************************

Dr Garib N Murshudov
Group Leader, MRC Laboratory of Molecular Biology
Hills Road 
Cambridge 
CB2 0QH UK
Email: ga...@mrc-lmb.cam.ac.uk 
Web http://www.mrc-lmb.cam.ac.uk




Reply via email to