and include a filter for blocking out UV light from the
excitation light within the optical path, unless DAPI or
CFP are being imaged.
(e) Autofocus: Due to the long timeline of these experiments,
there is bound to be drifting on the z-axis, which could
affect the focusing of the sample. Therefore, an autofocus
regime needs to be implemented, preferably on the phasechannel to avoid phototoxicity issues. For autofocusing
on the phase channel when imaging mycobacteria, we
typically scan 10 μm on the z-axis in 0.2-μm steps (see
Note 34).
4. Once all the parameters have been set and saved in the software,
image acquisition can start. The images are all saved automatically in the customized format of the acquisition software being
used, which is usually compatible with the most common
image analysis software (see Note 35).
5. Monitor the acquisition of the initial few time points, to make
sure the autofocus and other parameters are functioning
correctly.
6. Monitor the experiment on a regular basis to prevent potential
issues, at least every 2 h for fast-growing mycobacteria and at
least once a day for slow-growing mycobacteria.
3.9 Single-Cell Data
Analysis
and Processing Using
ImageJ Plug-Ins
and Macros
While there are several software packages that have been developed
and published for the automated segmentation and analysis of
time-lapse microscopy images of bacteria (see Note 10), these are
more suitable for bacteria such as E. coli or B. subtilis, where the
bacterial growth, shapes, and division are more amenable to
algorithm-based segmentation. Automated segmentation of mycobacterial cells and identification of cell division events is tricky
because the bacteria grow as long rods that tend to stick together
and branch, and because the events of cell division and cell separation are not simultaneous [36, 37]. Therefore, until now we have
been carrying out analysis of the time-lapse images by manual
segmentation using programs such as ImageJ or Fiji [47]. However,
the use of machine-learning approaches could offer valuable alternatives for automated analysis of single mycobacterial cells in space
and time in the coming future [50].
1. Import the image sequence stack into ImageJ or Fiji (see
Note 35).
2. Split the channels by using the “Substack Maker” or directly
importing into Fiji and rename each stack according to the
represented channel.
3. Use the function “Adjust ! Brightness/Contrast” for optimal
visualization of the cells to carry out diverse analytical procedures—single-cell segmentation; plot profile; cell counting
(Fig. 3a).
220
Giulia Manina and Neeraj Dhar
excitation light within the optical path, unless DAPI or
CFP are being imaged.
(e) Autofocus: Due to the long timeline of these experiments,
there is bound to be drifting on the z-axis, which could
affect the focusing of the sample. Therefore, an autofocus
regime needs to be implemented, preferably on the phasechannel to avoid phototoxicity issues. For autofocusing
on the phase channel when imaging mycobacteria, we
typically scan 10 μm on the z-axis in 0.2-μm steps (see
Note 34).
4. Once all the parameters have been set and saved in the software,
image acquisition can start. The images are all saved automatically in the customized format of the acquisition software being
used, which is usually compatible with the most common
image analysis software (see Note 35).
5. Monitor the acquisition of the initial few time points, to make
sure the autofocus and other parameters are functioning
correctly.
6. Monitor the experiment on a regular basis to prevent potential
issues, at least every 2 h for fast-growing mycobacteria and at
least once a day for slow-growing mycobacteria.
3.9 Single-Cell Data
Analysis
and Processing Using
ImageJ Plug-Ins
and Macros
While there are several software packages that have been developed
and published for the automated segmentation and analysis of
time-lapse microscopy images of bacteria (see Note 10), these are
more suitable for bacteria such as E. coli or B. subtilis, where the
bacterial growth, shapes, and division are more amenable to
algorithm-based segmentation. Automated segmentation of mycobacterial cells and identification of cell division events is tricky
because the bacteria grow as long rods that tend to stick together
and branch, and because the events of cell division and cell separation are not simultaneous [36, 37]. Therefore, until now we have
been carrying out analysis of the time-lapse images by manual
segmentation using programs such as ImageJ or Fiji [47]. However,
the use of machine-learning approaches could offer valuable alternatives for automated analysis of single mycobacterial cells in space
and time in the coming future [50].
1. Import the image sequence stack into ImageJ or Fiji (see
Note 35).
2. Split the channels by using the “Substack Maker” or directly
importing into Fiji and rename each stack according to the
represented channel.
3. Use the function “Adjust ! Brightness/Contrast” for optimal
visualization of the cells to carry out diverse analytical procedures—single-cell segmentation; plot profile; cell counting
(Fig. 3a).
220
Giulia Manina and Neeraj Dhar