Zebrafish Image quantification

Hi all, I have some zebrafish images where I have to quantify the GFP expression of heart,liver,kidney and brain altogether. I tried something like this:

a) Cropped the area of my interest (from head to trunk region).

b) subtracted background: rolling ball radius=50 pixels.

c) Min threshold value ranged from 30-100. Maximum threshold value was 255.

I am new at imageJ. Can anyone suggest me how to subtract background value and keep minimum threshold value same for all images OR am I doing it in right way?


Hi Moonlit,
I have a few questions to understand your problem. What kind of images do you have? Confocal 3D or 2D widefield images?
The steps that you are describing, are to create a masks which you want to use for measuring the intensity levels, right?
As you are saying that you want to measure the GFP signal, it probably varies between the images. So, you will have to use different thresholds or autothresholding. Do you have any reference marker for the structures you want to measure? Then you could eventually use a smaller range of threshold levels.


Hi, Sorry for being late. I am doing automatic imaging. Since I have to quantify multiple organs in same image, my threshold varies. I don’t have any reference marker. Is it the right way to quantify?


So with quantify GFP expression means: You want to analyze the “area” that is positive for GFP? I would say it is not an ideal quantification, but yeah in your case it could be the way to go. Have you tried some of the auto thresholding methods? I think by changing the threshold manually, you are introducing a bias even if you try to stay neutral. I usually try to do a “real” manual quantification with 3 example datasets and compare these vs. my workflow. If the results are similar you should be not so wrong.