Performs auto-segmentation of cell nuclei and measures the number and size of radiation induced DNA lessions within the nucleus.
In order to use this script, download Fiji and update it (Help > Update...
). After restarting Fiji, run the update command again and add StarDist to the update sites as shown here. Then, open the script by drag & dropping the file FociEval_V2020_2_JM.ijm
onto the Fiji toolbar. Click Run
to start the script.
A user interface will then pop up:
]
Parameters to be set and what they do:
Input image
: file path to a n-channel image, out of which one should be a DAPI staining and the other one showing the DNA lessions ("foci").Foci prominence
: Determines how sensitive the Foci detection should be. High numbers lead to fewer detected foci, low numbers yield more foci detections. It is highly encouraged to optimize this parameter before using this script. To do so, use theProcess > Find Maxima > ..
function from the Fiji toolbar, set the output type toPoint Selection
and check thePreview point selection
box. ou can then try different prominence value.Intensity cutoff
: Determines which part around a detected foci will be counted as part of the foci and which is not. If set to0.5
(which is suggested), this parameter corresponds to a full-width at half maximum area criterion.Pixel size
: Check if correct - to validate, open your image and pressCtrl + I
to inspect the metadata of your image.gH2AX
channel: Number of slice in the stack of channels that contains the image with the foci.DAPI
channel: Number of slice in the stack of channels that contains the DAPI staining. (Note: Start counting at 1)Minimal nucleus size
: Exclude nuclei smaller than this size from the analysisFoci brightness (lower percentile)
: The brightness of all foci in the image is measured and nuclei with a brightness lower than this value will be excluded from the analysisFoci brightness (upper percentile)
: The brightness of all foci in the image is measured and nuclei with a brightness higher than this value will be excluded from the analysis. It makes sense to use this parameter to exclude, i.e., dividing nuclei from the analysis.Boundary exclusion radius
: Cells this close to the edge of the image will be excluded from the analysis.Batch mode
: Hides the processing in the background if checked - also makes the analysis faster.Save cell wise images
: Saves a copy of the DAPI staining, the foci staining and the derived segmentations of the nucleus and the foci in a separate image for introspection.
This section describes in brief what the script does and how it works.
-
A sementation of the nuclei is first obtained with the StarDist plugin and some of the nuclei are excluded according to the above-described parameters. You are prompted with a review dialogue to approve.
-
The script iterates over every nucleus and detects the local maxima in the gH2AX-channel according to the prominence parameter above:
-
A tesselation of the nucleus with the foci positions as centers is then created to split the nucleus into separate foci-assigne dregions. The background (defined as lower 5% quartile) and the maximum intensity are then measured for every foci.
-
The size of the foci are measured by setting a threshold at
BG + 0.5 * ( Max - BG)
for each foci region, whereMax
is the intensity maximum of each foci andBG
is the local background around each foci.
If you use this tool for your analysis, please cite the following sources:
[1] Uwe Schmidt, Martin Weigert, Coleman Broaddus, and Gene Myers. Cell Detection with Star-convex Polygons. International Conference on Medical Image Computing and Computer-Assisted Intervention (MICCAI), Granada, Spain, September 2018.
[2] Rassamegevanon, T., Feindt, L., Koi, L., Müller, J., Freudenberg, R., Löck, S., Sihver, W., Çevik, E., Kühn, A. C., von Neubeck, C., Linge, A., Pietzsch, H. J., Kotzerke, J., Baumann, M., Krause, M., & Dietrich, A. (2021). Molecular Response to Combined Molecular- and External Radiotherapy in Head and Neck Squamous Cell Carcinoma (HNSCC). Cancers 2021, Vol. 13, Page 5595, 13(22), 5595. https://doi.org/10.3390/CANCERS13225595
[3] Suckert, T.; Rassamegevanon, T.; Müller, J.; Dietrich, A.; Graja, A.; Reiche, M.; Löck, S.; Krause, M.; Beyreuther, E.; von Neubeck, C. Applying Tissue Slice Culture in Cancer Research—Insights from Preclinical Proton Radiotherapy. Cancers 2020, 12, 1589. https://doi.org/10.3390/cancers12061589
- Stardist failing for larger images (red Java-specific error message): In this case, change the
nTiles
parameter in the stardist command to a higher value (i.e., 20). You can find the Stardist command by search forrun("Command From Macro", "command=[de.csbdresden.stardist.StarDist2D]"
in the script. Background: Before submitting the data through the Stardist algorithm, the image is cut into different tiles under the hood. The single tiles are smaller parts of the image and thus do not exceed the memory limits of the compiuter as the whole image being processed at once would do.