[phenixbb] sca file & anomalous difference map: understanding the story from the scratch,

Nathaniel Echols nechols at lbl.gov
Mon Jun 13 07:26:04 PDT 2011


On Mon, Jun 13, 2011 at 5:28 AM, Haytham Wahba <haytham_wahba at yahoo.com>wrote:

> second Q?
> using output.sca (with scale anomalous option on) for molecular Replacment
> generate MR.mtz
> which shows F,SIGF in data labels under phenix.refine
>
> the generated (refine_data.mtz) shows:
> I-obs(+),SIGI-obs(+),I-obs(-),SIGI-obs(-)
>
> what is going on in first and second run?
>

Two different things going on here:

1) in Phenix, the MR programs (AutoMR, or the Phaser GUI) will take any
format input (amplitudes or intensities, anomalous or non-anomalous), but
the output will always be non-anomalous amplitudes in MTZ format.  Note: do
not use MR.mtz for refinement!*  The data are corrected for anisotropy, and
thus are no longer the experimental observations.

2) When you use an anomalous Scalepack file for refinement, phenix.refine
will output the same data (unmodified) along with your newly generated
R-free flags in MTZ format, which is what you're seeing.  You should hold on
to this file and use it for all subsequent refinements, since you have now
refined against that particular set of R-free flags.

(A further note: the Phaser GUI also writes out a data file in MTZ format,
which contains the original data, but modified to be non-anomalous
amplitudes.  This is a Phaser quirk and can be ignored, but the data in that
file [which also ends in _data.mtz] are safe to use.)

third Q?
> i want to generate anomalous difference map using phenix.map
>
> but when i use output.sca (with scale anomalous option: ON) as reflections
> file
> the data label show ((i-obs,sigma))
>
>
Unfortunately Phenix does not always treat labels the same between different
reflection file formats.  In MTZ files there is a strict correspondence
between "column labels" and what you see in the various programs in Phenix,
so anomalous intensities with sigmas will always appear as something like
"I(+),SIGI(+),I(-),SIGI(-)".  Scalepack files will always appear as
"i_obs,sigma" whether or not they contain anomalous data.  If they are
genuinely unmerged (as Tom explained), they will appear as
"i_obs,sigma,merged", which is a signal that Phenix will merge the data
internally.

how to convert this SCA file (if it is merged) to unmerged sca file to creat
> anomalous difference map for protein contains Br, Sn and Sulpher??


You can use the refine_data.mtz file for this - you should use a file with
R-free flags anyway.  Alternately, use the French&Wilson program to convert
the intensities to amplitudes (while preserving the anomalous signal) with
correction for weak and negative values; it will also add R-free flags if
present.  However, if you have already run a round of refinement with the
anomalous output.sca as input, phenix.refine should have output an anomalous
difference map along with the other map coefficients, so there is no need to
make it a separate step.  (The GUI may not be automatically opening this map
in Coot, however - so you might need to choose the "Open MTZ..." option in
the first menu of Coot, and load the file manually.  I'll double-check this
when I get in to work.)

A general recommendation: if you now have an MTZ file with your anomalous
data and a set of R-free flags, start using this for everything you do in
Phenix, and ignore output.sca from now on - it will prevent confusion and
ensure that you don't end up with a new set of R-free flags by accident.

-Nat
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://phenix-online.org/pipermail/phenixbb/attachments/20110613/5f284e9c/attachment-0001.htm>


More information about the phenixbb mailing list