How to patch CCP4 cif2mtz so it can handle PDB entries with Map Coefficients.
Introduction
- Some recently released PDB entries, for instance 3u57 have structure factor files that contain map coefficients. The names of these items are defined in the PDB mmCIF exchange dictionary version 1.0697 that was released around June 2011, and follow REFMAC conventions:
| item | Description from http://mmcif.pdb.org/dictionaries/mmcif_pdbx.dic/Categories/refln.html |
| _refln.pdbx_DELFWT | The weighted structure factor amplitude for the mFo-DFc map. |
| _refln.pdbx_DELPHWT | Associated phase for the mFo-DFc map. |
| _refln.pdbx_FWT | The weighted structure factor amplitude for the 2mFo-DFc map. |
| _refln.pdbx_PHWT | Associated phase for the 2mFo-DFc map. |
- (We assume that it is normal to supply both the Fo-Fc and 2Fo-Fc map coefficients at the same time)
- The deposition and archiving of map coefficient is a really good idea because it means that final maps as interpreted by the crystallographer will then be available. In cases where map interpretation is important being able to see the original rather than third-party maps is valuable.
- To be able to visualise the deposited maps (eg in Coot) it is often necessary to convert the structure factor file to MTZ format using the CCP4 program cif2mtz
- but if we try to do this with the current version of cif2mtz the presence of the Map Coefficient data causes cif2mtz to terminate with an error.
- The Uppsala EDS server also currently has a problem with 3u57: 3u57 EDS page that may well have the same origin.
- This page shows how cif2mtz can be patched so that it can convert cif structure factor files containing the map coefficients.
- There is a separate page with some ideas on how to extend cif2mtz to handle new data types better in the future.
Patching cif2mtz to handle PDB entries with Map Coefficients
- These instructions are for a CCP4 6.2.0 installation from source that has been setup correctly.
- to be able to convert a cif structure factor files containing the map coefficients we need to do two things:
- patch cif2mtz.f, then compile and link it to get a new executable.
- produce a (compiled) mmCIF data dictionary that contains definitions for those additional items required added.
- It should be noted that the updated mmCIF data dictionary is still based on the CCP4 6.2.0 version (2.0.09, 8th March 2005) with the required additional items defined. It is NOT possible to use the latest PDB mmCIF Exchange Dictionary (available from http://mmcif.pdb.org/) because there are some modifications in CCP4 dictionary that are not present in the Exchange Dictionary.
Download the two patches cif2mtz_20111213.f.patch and cif_mm_20111213.dic.patch to somewhere, and apply them by doing
% cd $CCP4/src % cp -p cif2mtz.f cif2mtz.f_orig % patch cif2mtz.f /place/you/downloaded/cif2mtz_20111213.f.patch % cd $CCP4/lib/data % cp -p cif_mm.dic cif_mm.dic_orig % patch cif_mm.dic /place/you/downloaded/cif_mm_20111213.dic.patch
Compile and install the new version (this will replace cif2mtz and $MMCIFDIC with versions capable of handling the new terms):
% cd $CCP4/src % make cif2mtz % make instsome % cd $CCP4/lib/ccif % make cifdic_to_symtab % cd $CCP4/lib % cp cif_mmdic.lib cif_mmdic.lib_orig % ccif/cifdic_to_symtab data/cif_mm.dic cif_mmdic.lib 199
To test the extended capability, the following could be done:
% wget http://www.rcsb.org/pdb/files/r3U57sf.ent.gz % gunzip r3U57sf.ent.gz % echo END | cif2mtz hklin r3U57sf.ent hklout 3U57.mtz % mtzdmp 3U57.mtz | grep | awk '/Col Sort/,/STATISTICS/'
which should show
Col Sort Min Max Num % Mean Mean Resolution Type Column num order Missing complete abs. Low High label 1 ASC 0 43 0 100.00 15.9 15.9 65.16 2.43 H H 2 NONE 0 53 0 100.00 19.7 19.7 65.16 2.43 H K 3 NONE 0 89 0 100.00 33.3 33.3 65.16 2.43 H L 4 NONE 0.0 1.0 0 100.00 0.95 0.95 65.16 2.43 I FREE 5 NONE 60.2 7189.7 0 100.00 775.37 775.37 65.16 2.43 F FP 6 NONE 8.3 255.6 0 100.00 43.04 43.04 65.16 2.43 Q SIGFP 7 NONE 0.2124267.4 0 100.00 817.86 817.86 65.16 2.43 F FC 8 NONE 0.0 360.0 0 100.00 176.40 176.40 65.16 2.43 P PHIC 9 NONE 0.000 1.000 0 100.00 0.805 0.805 65.16 2.43 W FOM 10 NONE 0.0 4181.4 0 100.00 124.06 124.06 65.16 2.43 F DELFWT 11 NONE 0.0 360.0 0 100.00 169.21 169.21 65.16 2.43 P PHDELWT 12 NONE 0.0 10794.0 0 100.00 739.62 739.62 65.16 2.43 F FWT 13 NONE 0.0 360.0 0 100.00 179.77 179.77 65.16 2.43 P PHWT No. of reflections used in FILE STATISTICS 55223
- The file 3U57.mtz can also be used in coot to visualise the maps contained in the deposition. An interesting density feature can be found close to residue TRP 281 where a -7.2 sigma negative density peak is found. The residue appears to have been built in a favourable rotamer with two adjacent water molecules. If the two water molecules are deleted then the TRP can be fit happily into density with the NE2 atom forming a nice hydrogen bond to carbonyl O 378 (a user exercise).
- There is a separate page with some ideas on how to extend cif2mtz to handle new data types better in the future.
Page by Clemens Vonrhein, Oliver Smart and Peter Keller original version 21 Dec 2011. Address problems, corrections and clarifications to buster-develop@globalphasing.com