[sharp-discuss] sharp error: No column "FP" of type "F" found in file

Clemens Vonrhein vonrhein at globalphasing.com
Fri Nov 4 14:20:08 CET 2011


Hi Sabine,

On Fri, Nov 04, 2011 at 01:49:29PM +0100, Sabine Schneider wrote:
> I got an Fe peak dataset extending to about 4A. I also got a native 
> dataset to 2.6A and a crapy MR solution.

SOunds good.

> Symmetry seems to be P43212, 1mol/asu
> I want to use sharp to combine all these and hope to get interpretable 
> maps.

Yes.

> So I used CAD to add to put the native togehter with the peak data and 
> the phases from Phaser.

You can do that - but if you want to run autoSHARP you don't need to
create such a combined MTZ file. autoSAHRP does it for you
automaticaly anyway.

I would always recommend using reflection files as close to the
original data processing as possible:

 * my current favourite is using *.sca files (merged) for each dataset
   in autoSHARP: these can be written easily with SCALA as well (in
   case you come from a non-HKL package like autoPROC or MOSFLM).

   The advantage: when doing HA detection inside autoSHARP (using
   SHELXC/D) this file-format is expected. By giving them directly to
   autoSHARP, it doesn't need to convert back to intensities from an
   input MTZ file.

 * otherwise: just use the MTZ file after TRUNCATE/ctruncate

The important point here is: these files should always have the
correct column type for each column name. Remember, amplitudes in a
MTZ file ccan have any name you wnat - but need to have the column
type 'F' (for their sigmas it is 'Q'). See also

  http://www.ccp4.ac.uk/html/mtzformat.html#coltypes

> First I used the MR phases to find the iron (1 in 30kDa) in an anomalous 
> difference fourir map. - looks reasonable

Good. So you will feed this site to autoSHARP by creating a Fe.hatom
file with fractional coordinates like

  ATOM Fe 0.1234 0.2345 0.3456

in your sharpfiles/datafiles directory.

> So I fed autosharp (1 wavelength MIR) with the mtz from CAD 
> (ccp4-6.2.0).

And the *.hatom file, right?

> The first thing I noticed that despite beeing defined as 
> 'expert user', sharp does not automatically gives me the option 'using 
> external phases'.

autoSHARP doesn't: this is something for SHARP.

We have added an option to give autoSHARP also a PDB file (e.g. from
poor MR solution) in which case it will use SHARP (plus the external
phases from the model) to find HA sites etc. It works very well - as
expected, since that feature has been in SHARP for 10+ years ... some
people call this MR-SAD nowadays I think. Just that the button in
autoSHARP was missing ;-)

It will be in the upcoming SHAR/autoSHARPP release.

> This wasn't showing up either when re-running an old job, where I
> had used that option with a different mtz file in sharp of
> autosharp.

Yes: SHARP allows you to specify external phase information. autoSHARP
doesn't (yet).

> Since I didn't want to use the MR phases initially anyway, I just 
> ignored it and keep going.

Yes - if you found your Fe site and gave it to autoSHARP, it should
try and phase on that single one.

Be careful with single-site phasing though: some spacegroups can play
funny tricks with you (the score in the two hands is identical). Make
sure that that Fe site stays in, gives reasonable phasing power and
that the LLG residual maps are clean.

> But when I submitted the job to sharp it stopped with the
> 
> 
>        1.1.2 Columns
>   ERROR   : No column "FP" of type "F" found in file

When doing your 'mtzdmp': can you check that the column type is
correct?

Or maybe your MTZ file is not readable by the httpd (running the SHARP
server)? Double-check the permissions of the file.

> An old job of mine is running , also as I said the external-phase panel 
> is not visible.
> I upgraded Ubuntu 9.10 a while back to 10.4 - I haven't used sharp since 
> than.

The underlying Linux distro should have no impact ... unless: check
that the CCP4 configuraiton in

  /where/ever/sharp/machines/*/ccp4.setup

is correct and not pointing to some older (non-functional) CCp4
installation?

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)
***************************************************************


More information about the sharp-discuss mailing list