re [ccp4bb]: generate pdb header

[Kim Henrick <henrick@ebi.ac.uk>]

***  For details on how to be removed from this list visit the  ***
***    CCP4 home page http://www.dl.ac.uk/CCP/CCP4/main.html    ***



 ===========================
 re data_harvesting for CCP4
 ===========================
 
 The latest version of CCP4 will o/p mmcif files
 for several of the programs mosflm/scala/truncate/mlphare/restrain/refmac
 at the end of a refinement one can make a xx.tar.gz file of these files
 and submit these for deposition
 
 At this stage the atom coordinates are not o/p in mmcif format
 and would have to be submitted as a separate file (within a set of files).

 (see at end of this note about deposition to the ebi)
 

 It is not possible to convert an existing CCP4 logfile into
 harvest format.  Harvest files are only generated by using
 the latest version of CCP4.

 The semi-automatic tracking of data through CCP4 will be become
 easier when the next version of mosflm is released - this will
 impose a responsibilty of the user to start labelling data sets
 and the labels will be included into the first mtz file. Subsequent
 use of these mtz files will transfer labels to both 'harvest results
 files for each step' and to subsequent mtz file headers. It is however
 the users responsibilty to track each stage of a structure solution
 and carry out ultimate book keeping. The harvest files from each stage
 can then be accumulated into a compressed tar file and submitted for 
 simplified deposition. In addition - it hasnt yet been used in the
 real world and we dont know how it will mesh with different in-house
 practical use of different software. As it is common to say start with
 denzo, use CCP4 then cns then o then refmac - only experience will
 allow resolution of any practical difficulties.


 The work done at ESRF/BNL and other sync sites to add data labels
 to image headers will further smooth the flow of information from
 data collection to refinement by having mosflm and hkl2000 read
 the headers and transfer this to derived files. This is not yet
 in place.

 [for NMR structures the EBI is at the momment initiating a comprehensive
  harvest solution with the cooperation with the new CCPN and in particular
  with a workshop to held at the EBI on the 7-8Feb 2000]


 One can also use CNS and use the mmcif_deposition macro
 which was in
    .../cns_solve_0.9a/inputs/xtal_mmcif/deposit_mmcif.inp
 this only does refinement
    .../cns_solve_0.9a/modules/xtal/exportmmcifrefine  
    .../cns_solve_0.9a/modules/xtal/exportmmcifstruct
 
 there is also in pdb format (read by autodep)
  .../cns_solve_0.9a/inputs/xtal_pdbsubmission/xtal_pdbsubmission.inp
  .../cns_solve_0.9a/modules/xtal/pdbsubmission
 
 The CNS macros (and the equivalent XPLOR PDB macros) deal only 
 with information known for refinement. 

 Both CCP4 pdbset and the CNS/XPLOR PDB macros can generate PDB SEQRES
 records for sequence, although there are problems in some cases with
 different conformations and the enthusiam that the macros add all,
 including water to SEQRES.


 There is also info to be written out from hkl2000 (denzo/scalepack)
 although this isnt currently the same as CCP4 (mosflm/scala)
 

 As far as deposition - the ebi is still using PDB(BNL) legacy code in
 the form of autodep (autodep.ebi.ac.uk) as we in the final stages of
 developing a completely new submission system that can handle these
 files. We have not as yet modified autodep to read mmcif files - a
 function this software could never do.

 However, until the new system is ready, if you want to use CCP4 or cns
 mmcif o/p in your deposition, then please do the following:

1. Start an autodep submission, upload the files and email
pdbhelp@ebi.ac.uk  to say you have done this.

2. We will convert the files to autodep and allow you to continue the
submission.


Kim Henrick