Sent on behalf of Sacha Urzhumtsev:
Dear Richard,
I answer to the whole CCP4bb since the question was quite general and
the answer may
be interesting for other people.
As already mentioned by Bruno Klaholz, low-resolution reflections are not
only "a few reflections more" but they contain information
complementary to that in high-resolution reflections, in other words,
that is different to it.
Low-resolution reflections correspond to 'electron density waves'
of a large period. A right combination of low-resolution
structure factors gives the molecular envelope. Inversely, excluding
low-resolution reflections from the Fourier series calculation
artificially increases the map values in some large regions and decreases
it in others eventually destroying the molecular envelope in the maps.
If the peaks of the Fourier series are relatively low, this loss
of low-resolution data may completely kill the signal.
A map improvement by "phase extension for low-resolution" was illustrated
by Podjarny et al. (1981) Acta Cryst., A37, 662-668, and by Schevitz et al.
(1981) Acta Cryst., A37, 669-677. Later we studied in purpose the role
of low-resolution data for the map quality (1991; Acta Cryst., A47,
794-801).
In particular a practical illustration is given (EFG structure solution)
where an addition of 29 reflections to about 3000 available drastically
improved the map. At my knowledge, low-resolution data were crucial also
for the ribosomal structure solution (J.Cate).
Also as Bruno noted, the same data are important for refinement of
atomic models
(see for example Kostrewa, 1997, CCP4 Newslet., 34; I remember he was
presenting
something earlier but did not find the trace of that presentation). For
teaching
goals, recently with our students we prepared an example when a rigid-body
refinement of a protein model fails using 7377 reflections of the
resolution
2-20 A and gives a perfect answer when 12 (!) reflections only are added
at the resolution below 20 A.
Finally, the same low-resolution reflections may significantly
reinforce molecular-replacement translation searches (see
1995; Acta Cryst., D51, 888-895; Fokine & Urzhumtsev, 2002, Acta Cryst,
D58,
72-74).
Obviously, getting low-resolution data requires special efforts during
experiment
(see a recent discussion in the CCP4bb) but it is worthy to do and is
feasible in
practice at most of synchrotrons. Another feature of these data is a
very strong
contribution of the bulk solvent that should be taken into account as
soon as
one starts to use an atomic model.
I hope these few comments and some bibliography answer your questions and
may help in some studies.
With best regards,
Sacha
sa...@igbmc.fr
On 2/20/2009 8:22 AM, Richard Gillilan wrote:
Several times I have heard that low order (small angle) reflections
are more important when solving low-resolution structures. I presume
it is more than just a question of obtaining greater number of
reflections.
Does anyone know why low-order reflections are so important in these
cases?
Richard Gillilan
MacCHESS
