value at mouse position” is checked. Now, whenever you drag
your mouse on a point in the 3D FSC, the “radius” value
should be displayed in the dialog box in the bottom left-hand
corner of Chimera. To get the resolution at that particular
point, you simply have to divide the box size by that value (see
Note 17). It can also be convenient to display radial slices of
the 3D FSC. To do this, go to Tools ! Viewing Controls !
Side View and drag the clipping panels to constrain to the
center of the volume. The results of this procedure, for various
reconstructions, are displayed (see Figs. 3c and 7).
7. Rotate the map. You should see variations in the Z-resolution
relative to the viewing angle. If the map is anisotropic, the Zresolution will change, as will the color of the map (see Fig. 4
and Notes 15 and 16).
3.3 Analyses
of the Individual
Results
Here, we will elaborate on the results of the operations above and
comment on analyzing individual case studies. The different case
studies show variations in both local and directional resolution.
3.3.1 Half-Map
and Map-Model FSCs
The first step in evaluating map quality is computing global halfmap, and optionally (if a model is previously available or has been
derived from the data), map-model FSCs. A half-map FSC
describes the internal consistency between two reconstructions
computed from half-sets of the data, while a map-model FSC
describes the consistency between an experimental map and an
atomic model (see Note 11). These steps are standard in every
refinement program and have not been described in the current
protocol, but we provide the results of these operations for reference and as examples distinguished by different types of anisotropies. A suitable mask was applied to the experimental maps and to
the maps generated from the atomic models (see Note 12). The
AAV2 capsid has icosahedral symmetry. The 60-fold averaging not
only boosts the signal-to-noise ratio, but also results in uniform
distribution of projection views (it also improves the sampling
compensation factor, as described recently [18]. As a result, the
reconstruction is isotropic, and a smooth half-map FSC curve
indicates a healthy reconstruction (see Fig. 1a). The half-map and
map-model nominal resolution values are also similar to one
another. In contrast, the HA trimer is anisotropic, and the true
“top” projection views predominate over other projection views. A
comparison of two HA trimer datasets was performed to demonstrate the effects of directional resolution anisotropy: one dataset
was collected without tilting (0
) and one dataset was collected
with a 40
tilt angle of the specimen stage [6]. The half-map FSC
curve from the 0
-dataset shows an apparent resolution of 3.3 Å (see
Fig. 1b, blue). However, the map-model FSC curve (using an
externally derived and previously available atomic model derived
Local and Directional Resolution in Cryo-EM Maps
173
your mouse on a point in the 3D FSC, the “radius” value
should be displayed in the dialog box in the bottom left-hand
corner of Chimera. To get the resolution at that particular
point, you simply have to divide the box size by that value (see
Note 17). It can also be convenient to display radial slices of
the 3D FSC. To do this, go to Tools ! Viewing Controls !
Side View and drag the clipping panels to constrain to the
center of the volume. The results of this procedure, for various
reconstructions, are displayed (see Figs. 3c and 7).
7. Rotate the map. You should see variations in the Z-resolution
relative to the viewing angle. If the map is anisotropic, the Zresolution will change, as will the color of the map (see Fig. 4
and Notes 15 and 16).
3.3 Analyses
of the Individual
Results
Here, we will elaborate on the results of the operations above and
comment on analyzing individual case studies. The different case
studies show variations in both local and directional resolution.
3.3.1 Half-Map
and Map-Model FSCs
The first step in evaluating map quality is computing global halfmap, and optionally (if a model is previously available or has been
derived from the data), map-model FSCs. A half-map FSC
describes the internal consistency between two reconstructions
computed from half-sets of the data, while a map-model FSC
describes the consistency between an experimental map and an
atomic model (see Note 11). These steps are standard in every
refinement program and have not been described in the current
protocol, but we provide the results of these operations for reference and as examples distinguished by different types of anisotropies. A suitable mask was applied to the experimental maps and to
the maps generated from the atomic models (see Note 12). The
AAV2 capsid has icosahedral symmetry. The 60-fold averaging not
only boosts the signal-to-noise ratio, but also results in uniform
distribution of projection views (it also improves the sampling
compensation factor, as described recently [18]. As a result, the
reconstruction is isotropic, and a smooth half-map FSC curve
indicates a healthy reconstruction (see Fig. 1a). The half-map and
map-model nominal resolution values are also similar to one
another. In contrast, the HA trimer is anisotropic, and the true
“top” projection views predominate over other projection views. A
comparison of two HA trimer datasets was performed to demonstrate the effects of directional resolution anisotropy: one dataset
was collected without tilting (0
) and one dataset was collected
with a 40
tilt angle of the specimen stage [6]. The half-map FSC
curve from the 0
-dataset shows an apparent resolution of 3.3 Å (see
Fig. 1b, blue). However, the map-model FSC curve (using an
externally derived and previously available atomic model derived
Local and Directional Resolution in Cryo-EM Maps
173