If there are missing points: Select “Track with Fiducial Model
as Seed”. This will use the existing fiducial model as a seed and
often allows more of the beads to be tracked. Repeat this tracking
with the fiducial model as the seed until the number of missing
points no longer decreases. If there are still missing points, select
“Fix Fiducial Model”.
On the bead fixer window, select “Autocenter” and input the
diameter of your fiducial. Select “Go to Next Gap”. Use the page
up/page down keys to scroll between images and follow one bead
through the tilt-series. Middle-click on the center of the bead to
add model points to the images where fiducial tracking failed. Press
“S” on the keyboard to save the modified model file (see Note 16).
Select “Track with Fiducial Model as Seed” again and select
“Done”.
3.7.4 Fine Alignment
Select “Compute Alignment” and check the residual error mean
and standard deviation in the log window. A mean near 0.5 is ideal,
but not routine for thicker samples; typically, the mean fluctuates
between 0.5 and 1.5. For details on improving the residual error,
consult the IMOD users guide at the URL listed above. Select
“Done”.
3.7.5 Tomogram
Positioning
For neuronal branches, use a sample tomogram thickness of ~800.
Select “Create Whole Tomogram” with a binning of three and then
select “Create Boundary Model.” Rotate the tomogram 90
under
the menu “Edit ! Image ! Flip/Rotate”. Define the upper and
lower boundaries of the cell by drawing a horizontal line above and
below the region of interest near one of its edges. Lines are drawn
by creating two model points with the middle mouse button, and
they should be automatically connected. Do the same at the center
and far end of the object. Save the model, and select “Create Final
Alignment”.
3.7.6 Final Aligned Stack
Determine which binning/pixel size allows for the best contrast
without losing the desired detail. To increase contrast and make
future segmentation calculations faster, bin the tilt-series another
two-fold to create a final aligned stack that is 1 k  1 k. To set the
binning to two-fold, change the “Aligned Image Stack Binning”
parameter to 2. Select “Create Full Aligned Stack” and then view
the stack to make sure the alignment is satisfactory. Select “Done”.
3.7.7 Tomogram
Generation
If using parallel processing, ensure that “Parallel Processing” is
checked, and then select the # of CPUs to use. If using a CUDAenabled GPU, select “GPU Processing”. If a boundary model was
generated in the “Tomogram Positioning” section, the “Tomogram Thickness in Z” parameter will already be set. If a boundary
model was not generated, set it to a relatively large number of pixels
Cryotomography of Neurons
39
as Seed”. This will use the existing fiducial model as a seed and
often allows more of the beads to be tracked. Repeat this tracking
with the fiducial model as the seed until the number of missing
points no longer decreases. If there are still missing points, select
“Fix Fiducial Model”.
On the bead fixer window, select “Autocenter” and input the
diameter of your fiducial. Select “Go to Next Gap”. Use the page
up/page down keys to scroll between images and follow one bead
through the tilt-series. Middle-click on the center of the bead to
add model points to the images where fiducial tracking failed. Press
“S” on the keyboard to save the modified model file (see Note 16).
Select “Track with Fiducial Model as Seed” again and select
“Done”.
3.7.4 Fine Alignment
Select “Compute Alignment” and check the residual error mean
and standard deviation in the log window. A mean near 0.5 is ideal,
but not routine for thicker samples; typically, the mean fluctuates
between 0.5 and 1.5. For details on improving the residual error,
consult the IMOD users guide at the URL listed above. Select
“Done”.
3.7.5 Tomogram
Positioning
For neuronal branches, use a sample tomogram thickness of ~800.
Select “Create Whole Tomogram” with a binning of three and then
select “Create Boundary Model.” Rotate the tomogram 90
under
the menu “Edit ! Image ! Flip/Rotate”. Define the upper and
lower boundaries of the cell by drawing a horizontal line above and
below the region of interest near one of its edges. Lines are drawn
by creating two model points with the middle mouse button, and
they should be automatically connected. Do the same at the center
and far end of the object. Save the model, and select “Create Final
Alignment”.
3.7.6 Final Aligned Stack
Determine which binning/pixel size allows for the best contrast
without losing the desired detail. To increase contrast and make
future segmentation calculations faster, bin the tilt-series another
two-fold to create a final aligned stack that is 1 k  1 k. To set the
binning to two-fold, change the “Aligned Image Stack Binning”
parameter to 2. Select “Create Full Aligned Stack” and then view
the stack to make sure the alignment is satisfactory. Select “Done”.
3.7.7 Tomogram
Generation
If using parallel processing, ensure that “Parallel Processing” is
checked, and then select the # of CPUs to use. If using a CUDAenabled GPU, select “GPU Processing”. If a boundary model was
generated in the “Tomogram Positioning” section, the “Tomogram Thickness in Z” parameter will already be set. If a boundary
model was not generated, set it to a relatively large number of pixels
Cryotomography of Neurons
39