single map as input. The resulting files from all these programs
can generally be analyzed using a similar procedure as that
described in the current protocol.
2. The mask constrains the analysis to specific regions of a map
and, by means of excluding regions of solvent outside of the
density, provides a more accurate measure of global resolution.
In Sparx, this mask will also define the region to which an
optional local filter will be applied (see Note 6). The mask can
be generated using most processing packages.
3. Sharp or hard edges from the mask can skew and/or artificially
inflate the local resolution within specific regions of the map.
The amount of inflation can be measured using the highresolution noise substitution test [34]. Care must be taken
when generating the mask to make sure that it maintains soft
edges and does not cut into the density. Each package has its
own protocol to generate a mask, but generally, the procedures
are: binarize the experimental map at a defined threshold !
extend the mask ! soften the edges.
4. A different threshold (e.g., 0.5) can be selected for computing
the nominal resolution value at each voxel centroid.
5. For more usages of sxlocres.py, you can use the command
sxlocres.py -help or check the website:
http://sparx-em.org/sparxwiki/sxlocres
6. The output local resolution map can also be used to locally
filter the input volume. This step is not covered in the current
protocol, but can be optionally performed.
7. In practice, identifying the best resolution range by which to
color the input map involves some trial and error and depends
on what the user wants to highlight. The user can usually
assume that the distribution of local resolution values is centered around the global resolution of the map (although this
need not be the case, e.g., for dynamic regions) and select a set
of resolution values that color the volume with, e.g., ~0.5 Å
intervals across the resolution range. Chimera provides a fivecolor palette as default; further adjustments of the resolution
intervals can be performed, with the general goal of displaying
the full range of colors. The range can also be estimated by
looking at a histogram of the local resolution voxel values,
which is displayed in Chimera. Sometimes, the range of local
resolution values is very narrow (e.g., this is the case with the
AAV2 example that is described in the current protocol, see
Fig. 2a). In this case, the user should decrease the resolution
interval range. Other times, especially when there is substantial
heterogeneity (conformational or compositional) in the complex, the range could be larger (e.g., this is the case with the
SgrAI helical filament described in the current protocol, see
180
Sriram Aiyer et al.
can generally be analyzed using a similar procedure as that
described in the current protocol.
2. The mask constrains the analysis to specific regions of a map
and, by means of excluding regions of solvent outside of the
density, provides a more accurate measure of global resolution.
In Sparx, this mask will also define the region to which an
optional local filter will be applied (see Note 6). The mask can
be generated using most processing packages.
3. Sharp or hard edges from the mask can skew and/or artificially
inflate the local resolution within specific regions of the map.
The amount of inflation can be measured using the highresolution noise substitution test [34]. Care must be taken
when generating the mask to make sure that it maintains soft
edges and does not cut into the density. Each package has its
own protocol to generate a mask, but generally, the procedures
are: binarize the experimental map at a defined threshold !
extend the mask ! soften the edges.
4. A different threshold (e.g., 0.5) can be selected for computing
the nominal resolution value at each voxel centroid.
5. For more usages of sxlocres.py, you can use the command
sxlocres.py -help or check the website:
http://sparx-em.org/sparxwiki/sxlocres
6. The output local resolution map can also be used to locally
filter the input volume. This step is not covered in the current
protocol, but can be optionally performed.
7. In practice, identifying the best resolution range by which to
color the input map involves some trial and error and depends
on what the user wants to highlight. The user can usually
assume that the distribution of local resolution values is centered around the global resolution of the map (although this
need not be the case, e.g., for dynamic regions) and select a set
of resolution values that color the volume with, e.g., ~0.5 Å
intervals across the resolution range. Chimera provides a fivecolor palette as default; further adjustments of the resolution
intervals can be performed, with the general goal of displaying
the full range of colors. The range can also be estimated by
looking at a histogram of the local resolution voxel values,
which is displayed in Chimera. Sometimes, the range of local
resolution values is very narrow (e.g., this is the case with the
AAV2 example that is described in the current protocol, see
Fig. 2a). In this case, the user should decrease the resolution
interval range. Other times, especially when there is substantial
heterogeneity (conformational or compositional) in the complex, the range could be larger (e.g., this is the case with the
SgrAI helical filament described in the current protocol, see
180
Sriram Aiyer et al.