¶ Unibiased Measurements
¶ BIOP recommandation
Do not use the same channel for segmenting the image and measuring intensities.
To avoid bias in your measurements, make sure to have a detection that is independent
of your variable of interest
An example of common bias is show on this OMERO.figure.
Here is the script used to make the processing. The image can be downloaded from OMERO
originalTitle = getTitle()
Stack.setPosition(1, 1, 1);
run("Duplicate...", " channels=1");
rename("DAPI")
run("Blue");
run("Enhance Contrast", "saturated=0.35");
dapiTitle = getTitle()
selectImage(originalTitle);
run("Duplicate...", " channels=1");
rename("DAPI-EdU-ROIs")
run("Blue");
run("Enhance Contrast", "saturated=0.35");
dapiEduTitle = getTitle()
selectImage(originalTitle);
run("Duplicate...", " channels=1");
rename("DAPI-mask")
dapiMaskTitle = getTitle()
selectImage(originalTitle);
Stack.setPosition(2, 1, 1);
run("Duplicate...", " channels=2");
rename("Edu")
run("Green");
run("Enhance Contrast", "saturated=0.35");
eduTitle = getTitle()
selectImage(originalTitle);
run("Duplicate...", " channels=2");
rename("Edu mask")
run("Green");
run("Enhance Contrast", "saturated=0.35");
eduMaskTitle = getTitle()
selectImage(originalTitle);
run("Duplicate...", " channels=2");
rename("Edu-DAPI-ROIs mask")
run("Green");
run("Enhance Contrast", "saturated=0.35");
eduDapiTitle = getTitle()
selectImage(dapiMaskTitle);
run("Gaussian Blur...", "sigma=2");
setAutoThreshold("Li dark");
run("Convert to Mask");
run("Analyze Particles...", "size=50.00-Infinity display exclude clear add");
selectImage(dapiTitle);
run("From ROI Manager");
selectImage(eduDapiTitle);
run("From ROI Manager");
selectImage(eduMaskTitle);
run("Gaussian Blur...", "sigma=2");
setAutoThreshold("Li dark");
run("Convert to Mask");
run("Analyze Particles...", "size=50.00-Infinity display exclude clear add");
selectImage(eduTitle);
run("From ROI Manager");
selectImage(dapiEduTitle);
run("From ROI Manager");
run("Tile");
From the above example, let's display the intensiy measurements in a dot plot to compare both conditions. We can fortunatly set a positive/negataive cut-off at 20. The proposrtions of positive/total are, respectively
- 7/30 vs 7/21
- 23% vs 33%
