Volume pre-processing

When preparing to match two volumes from different sources, there are a number of pre-processing steps that help the comparison:

Bandpass filtering

The resolution range of the higher resolution volume will typically be restricted to that of the lower resolution volume, since the higher resolution information cannot be matched.

Background peak

The distribution of density values in the map typically has a peak representing the large number of voxels having a background value. The map values are all shifted by a constant amount so that this peak appears at zero.

Contour level

The map contains a continuous range of values at its grid positions. While high values are expected to represent the macromolecule and low values the surrounding solvent, there is no explicit division between the two. When interpreting a map, software will often set a contour level that represents the boundary between molecule and solvent, but this choice is somewhat subjective. Typically, it is chosen so that the enclosed volume corresponds to the known molecular weight at a typical molecular density.

Noise removal

Maps often contain many small peaks arising from noise in the data, which can confuse interpretation. These can be removed by “dusting”, i.e. by removing disconnected volumes smaller than a given cutoff. It may also be useful to hide dust in difference maps, in order to focus on major changes in the volume data.