Correlation of map and model after adjusting model for origin shifts with get_cc_mtz_pdb

Author(s)

Purpose

get_cc_mtz_pdb is a command line tool for adjusting the origin of a PDB file using space-group symmetry so that the PDB file superimposes on a map, obtaining the correlation of model and map, and analyzing the correlation for each residue.

Usage

How get_cc_mtz_pdb works:

get_cc_mtz_pdb calculates a model map based on the supplied PDB file, then uses RESOLVE to find the origin shift (using space-group symmetry) that maximizes the correlation of this model map with a map calculated with the supplied map coefficients in an mtz file. This shift is applied to the atoms in the PDB file to create offset.pdb and then the correlation, residue-by-residue of offset.pdb with the map is analyzed. Atoms and residues that are out of density or are in weak density are flagged.

You can set several parameters to define how the correlations are calculated.

By default model density is calculated using the atom types, occupancies and isotropic thermal factors (B-values) supplied in the PDB file. If you specify

scale=True

then an overall B as well as an increment in B-values for each atom beyond CB (for proteins) will be added to the values in the PDB file, after adjusting these parameters to maximize the map correlation.

If you specify

use_only_refl_present_in_mtz=True

then the model-based map will be calculated using the same set of reflections as the map calculated from your input mtz file. This reduces the contribution of missing reflections on the calculation (but the correlation is no longer the actual map-model correlation).

In the calculation of the map correlation in the region of the model, the region where the model is located is defined as all points within a distance rad_max of an atom in the model. The value of rad_max is adjusted in each case to maximize this correlation. Its value is typically similar to the high-resolution limit of the map.

Output files from get_cc_mtz_pdb

offset.pdb: A PDB file offset to match the origin in the mtz file.

Examples

Standard run of get_cc_mtz_pdb:

Running the get_cc_mtz_pdb is easy. From the command-line you can type:

phenix.get_cc_mtz_pdb map_coeffs.mtz coords.pdb

If you want (or need) to specify the column names from your mtz file, you will need to tell get_cc_mtz_pdb what FP and PHIB (and optionally FOM) are, in this format:

phenix.get_cc_mtz_pdb map_coeffs.mtz coords.pdb \
labin="FP=2FOFCWT PHIB=PH2FOFCWT"

Possible Problems

Specific limitations and problems:

The option to use atom selections in get_cc_mtz_pdb can cause some confusion, because the CC values obtained can depend on the atom selections. The way this works is that the PDB file (after atom selections) is used to calculate model density. This model density is compared to the map from your input map coefficients in the region surrounding the atoms selected. This model density can depend on the presence of nearby atoms (where the density extends a ways away from those atoms).

In versions of PHENIX up to 1.3-final, defaults were set to maximize the correlation coefficient rather than to give the correlation using the existing thermal parameters and including only the reflections present in the mtz file. These previous defaults were equivalent to using the values:

scale=True
use_only_refl_present_in_mtz=True

These defaults were changed so that the correlation values obtained by default in a case where no origin shifts are needed would correspond to those obtained by simply calculating (1) a map using the input map coefficients and (2) a map from the PBB file and then determining the correlation between these maps.

List of all available keywords