How to patch CCP4 cif2mtz so it can handle PDB entries with Map Coefficients

From: Oliver Smart
Date: 21 December 2011 14:40

Some recently released PDB entries (for instance 3u57) have structure factor files that contain map coefficients. We think this is a really good idea because it means that final maps as interpreted by the crystallographer will then be archived and available.

But there is a problem converting such files to mtz with the current cif2mtz program; indeed we raised this on the bb in September:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=CCP4BB;6be4044e.1109

To get around this we produced a quick-fix patch for cif2mtz that is available at:

http://www.globalphasing.com/buster/wiki/index.cgi?CCP4cif2mtzPatchToHandleMapCoefficients


Regards,

Oliver

| Dr Oliver Smart                             |
| Global Phasing Ltd., Cambridge UK           |
| http://www.globalphasing.com/people/osmart/ |

----------
From: <martyn.winn


Thanks for this. I have recently updated the CCP4-supplied dictionary to include standard mmCIF, exchange dictionary, plus extra CCP4 items. This will be in the next release.
I was about to update cif2mtz to recognise these extra columns. Will look at what you have done, may save me a job.

cif2mtz is based on the ccif library, which complains if you try to load a cif file which contains a non-dictionary item. I'm sure in the past I tried to find a way round this, and failed.

From the user point of view, the mmcif file is ascii and you can hand-edit the column item names. This is the get-out method if cif2mtz fails. Edit the item name to something that is recognised.

Cheers
Martyn

--
***********************************************************************
*                                                                     *
*               Dr. Martyn Winn    

----------
From: Peter Keller


Hi Martyn, This can be done, at least to the extent that the non-dictionary item can be ignored by the application rather than causing it to crash. (Reading data values from the mmCIF does absolutely require the item to have been defined in the dictionary though.) If you look at our suggested patch to cif2mtz to allow reading the map coefficient data you will get a hint about how to do this, or I can send details to you off-list or to ccp4-dev. This is not ideal (says he, not having forgotten his own involvement here :-)), and it is possible to imagine cases where this is not trivial to do because of conflicting requirements of different dictionaries. That is not intrinsic to the way that cif2mtz is written though, and could be fixed.

Regards,
Peter.

--
Peter Keller