[buster-discuss] buster maps in pymol

[Clemens Vonrhein]

Hi Donald,

On Tue, Sep 27, 2011 at 04:22:16PM -0400, Donald Damian Raymond wrote:
> I am having the same issue opening an MTZ generated by buster in the
> latest version of PyMOL.

Can pymol now open reflection files directly? I thought it could only
open maps ... but haven't used it for a while now.
 
> In refmac where is an option to generate a map around the atoms in
> the PDB that will ensure that the map is only generated around atoms
> in the pdb.

I guess you mean the option in the CCP4i interface to refmac, right?
refmac itself doesn't generate maps - it writes map coefficients in
the reflection MTZ file. The map calculation is then done with the
'fft' program followed by 'mapmask' to get a map covering a PDB file.

We do the same in BUSTER, just that the default is slightly different:
if you want the final maps already created by BUSTER to cover the PDB
file:

  % refine FinalMapsCoverPdb=yes ...

> It appears that the MTZ generated by buster only draws the map in
> the unit cell when loaded in PyMOL.

Ok, just to ensure that there is no confusion here (sorry if that is
too basic/detailed, but in case someone will stumble across that email
later): the MTZ file contains reflections (reciprocal space) and a map
is a real space electron density. Going from one to another involves a
FFT. For both there are definitions of asymmetric unit - and the
reflection MTZ file contains the reciprocal space asymmetric unit.

When calculating a map from that MTZ file, it will also generate
electron density in the real-space asymmetric unit - or (in case of
CCP4's 'fft' program in default mode) for the whole unit cell.

Most molecular display programs like Coot, O etc know about crystal
symmetry (eg. from the CRYST1 card of your PDB file) and can show you
density anywhere in space: as long as you have density for the
asymmetric unit you can always apply symmetry operators and cell
shifts.

Pymol doesn't seem to have gained that feature yet. So it is the users
task to generate a map covering the range they're interested in.

> However, in many cases some of the modeled atoms are not in the unit
> cell. See attached image.
> 
> It may be both a buster and PyMOL issue.

Yes, it's a map preparation issue triggered by a lack of a PyMOL
feature.

Just add the above argument to your 'refine' command (or add it to
your ~/.autoBUSTER file). Or run

  echo "LABI F1=FOFCWT PHI=PHFOFCWT" | \
    fft hklin refine.mtz mapout Fo-Fc.map
  echo "LABI F1=2FOFCWT PHI=2PHFOFCWT" | \
    fft hklin refine.mtz mapout 2Fo-Fc.map

to get the maps from the MTZ file. And then

  echo "BORDER 5.0" | \
    mapmask mapin Fo-Fc.map xyzin refine.pdb mapout Fo-Fc_cover.map
  echo "BORDER 5.0" | \
    mapmask mapin 2Fo-Fc.map xyzin refine.pdb mapout 2Fo-Fc_cover.map

to get maps with 5A border around the refine.pdb file.

However, a display program for crystal structures and density that
doesn't know about crystal symmetry and spacegroups seems to be
lacking something rather basic I think ;-)

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