Hi Tassos,
On Mon, Apr 12, 2010 at 04:10:57PM +0200, Anastassis Perrakis wrote:
> As for Frank's argument, I fully agree low resolution cases are a
> different ball game
> (talking about ball games, I usually agree with Frank, he is much
> bigger than me, and its healthier to agree with him, since I play
> football with him one every two years at the Gordon).
I agree too ... especially since big Frank is sitting 2 metres away
from me right now ;-)
> A 4 A case, or anything below 3.0-3.2 A, warrants very careful
> documentation of how maps were calculated, since its the quality of
> the experimental maps rather than quality of the refinement that
> will likely dominate model quality.
In those cases I find an overall value always less meaningful than a
more detailed break-down versus resolution (similar to SHELXC output -
which would be my preferential stat to be reported for SHELXC/D/E
solved structures). Often, a 3A dataset might only have meaningful
heavy atom phases to 4-5A - and then the whole procedure for
calculating maps, density modification, NCS averaging, automatic
building etc becomes quite impoprtant. After all, a missing domain in
those cases might just be due to very poor initial heavy atom phases
that were misleading density modification procedures into flattening
it away. So a good description of the procedure is even better (even
if it is only in supplementary material) than just a number in some
table 1.
I like the phasing power statistics as three numbers:
- what is the value in the lowest resolution range
It often shows issues with overloads, poor beamstops etc - which
can be a big problem later on.
- at what resolution does it drop below 1.0
This gives me an idea what the initial, purely heavy-atom phased
map will look like (and how one can improve the phases to the full
resolution of the data).
- overall value
Not very useful, I find.
> For higher resolution were nearly complete models are autobuilt from
> the modified maps, i find the reporting of anything more than the
> FOM (to satisfy my curiosity) a bit of a nuisance.
Yes (although I still like the phasing power better - but this looks
quite different for different phasing programs I guess). As long as it
isn't the FOM after density modification which can be highly
overestimated.
Cheers
Clemens
--
***************************************************************
* Clemens Vonrhein, Ph.D. vonrhein AT GlobalPhasing DOT com
*
* Global Phasing Ltd.
* Sheraton House, Castle Park
* Cambridge CB3 0AX, UK
*--------------------------------------------------------------
* BUSTER Development Group (http://www.globalphasing.com)
***************************************************************
|