Scaled unmerged data files

From CCP4 wiki
Jump to navigationJump to search


A deposition to the wwPDB may not only include merged reflection data (i.e. the typical input to the refinement program), but also scaled unmerged data (as part of the Improved support for extended PDBx/mmCIF structure factor files announced in 2021). A PDBx/mmCIF reflection data file with such data can be

  • generated by a MX software suite directly (which is often the preferred way, since it will typically also contain the data quality metrics describing the data) - see e.g. autoPROC from Global Phasing,
  • generated by this web app or the equivalent command-line program, both developed as part of project GEMMI.

The prerequisite for the latter is a file with scaled unmerged data.

Please remember that a full deposition of reflection data will usually contain

  • The data that went into the final refinement step (to provide the possibility for validation and re-refinement)
  • The output of the refinement program - including map coefficients for the 2mFo-DFc (electron) density map and the mFo-DFc difference density map. This will allow a user to compute the maps exactly as seen by the depositor, i.e. those maps any original model interpretation were based upon.
  • (optionally) the scaled and merged data before a final step of data selection, application of a cut-off or re-scaling.
  • The scaled and unmerged data discussed on this page.
  • (optionally) Additional reflection data that describe for example subsets of the data, different wavelengths or such.
  • A set of data quality metrics for each of those separate reflection datablocks (R-values, completeness, redundancy, CC1/2, mean I/sig(I) etc).

Where to find scaled unmerged data?


If you run Aimless from the command-line add option OUTPUT MTZ UNMERGED.


Scaled unmerged data is stored in scaled.mtz as well as in mmCIF file (GPhL developers recommend to use the mmCIF file and aB_deposition_combine to prepare files for deposition).


i1 has a few tasks that scale data:

  • Scale and Merge Intensities – it uses obsolete SCALA program. To write scaled unmerged data check ☑ Customize Scala process and select Output scaled unmerged & no outliers (?).
  • Symmetry, Scale, Merge (Aimless) – check ☑ Customize output options and select MTZ Output both merged and unmerged.
  • Find Symmetry, Scale & Merge (Scala) – obsolete, use Aimless instead.


To store scaled unmerged data in Data reduction - AIMLESS task you need to open Important Options and check:

output unmerged data as well as merged

CCP4 Cloud

Saving scaled unmerged data is not supported?


To output scaled unmerged MTZ file either add option unmerged_mtz=FILE:

dials.scale symmetrized.expt symmetrized.refl unmerged_mtz=scaled_unmerged.mtz

or use dials.export afterwards:

dials.export scaled.refl scaled.expt


The DIALS User Interface has 8 steps. The last one is export which writes unmerged MTZ file. The option Output Scaled Intensities must be checked (default) to write scaled unmerged data.


The 'no merge original index' macro for SCALEPACK produces a non-standard .sca file. It cannot be used, for example, for structure solution in HKL2000, and therefore gives a warning.


iMosflm can scale data with Aimless. Scaled unmerged MTZ file is written by default since v7.4.0 (soon to be released). Older versions had option (off by default) to write scaled unmerged data in the SCALEPACK format.


CCP4 SCALA program got obsoleted in favor of Aimless. It has option OUTPUT UNMERGED.




Scaled unmerged data from XDS CORRECT and XSCALE is written in the XDS_ASCII.HKL format. The GEMMI converter can use these files directly.


Since xdsme uses XDS/XSCALE, see above.


Scaled unmerged MTZ file is stored as DataFiles/*_scaled_unmerged.mtz.