Calculate the area in ISH slides

I am trying to analyze some ISH images but I am not sure how to get the total area with a positive signal for my target-specific probe. I am looking for CXCL13 in lung tissue. Any help would be welcome!

Hi @Nancy

Would you be able to post an example image? That would help us help you. And just indicate what you consider positive signal…

eta :slight_smile:

Hi @etadobson
Here is one of those images. I need to calculate the area occupied by the dark dots compared to the total area. Thank you for your help!

Hi @Nancy

So… I just opened your image in ImageJ… and I’m worried about a few things:

  1. Your image is a JPEG. Ideally - you should not be saving images as JPEGs - due to lossy compression - you can read more about that here. You should never use JPEG - especially for quantification analysis. It’s better to use TIFF or the original file format of your imaging system.

  2. If you zoom in on your image - you will see that the resolution of your dark spots is not high enough to measure reliably. Pixel-sized dark spots are abundant… so there is no way to say that they are real or noise or what? At least - I’d not consider this resolution reliable (perhaps others on the Forum here would disagree?) Have a look at the image below that I took of yours zoomed in. As it is almost impossible within this image to delineate your individual cells - I would doubt you could reliably measure your CXCL13 staining within.

All-in-all … would it be possible to re-acquire these images at a higher resolution and saving as TIFFs??

I know this is perhaps not the answer you’d wanted… sorry for that. But it’s the best way to get the best data you can.



Hi @etadobson
Thanks for your comments. I have the images in .tiff and I posted one for you here (20X magnification). I have the images with a high quality because they were acquired using the nanozoomer but I need to figure out how to do the analysis using imageJ.
I would like to ask you what is your suggestion for doing this analysis? I was trying to use imageJ with best quality but it seems like the numbers that I got correspond to the intensity of pixels but no the area corresponding to the specific staining.

I really appreciate your help!


I would suggest giving Trainable Weka Segmentation a try. I just used the default settings and drew a few traces and got a decent segmentation:

That should get you started to then do the area measurements you need.

eta :slight_smile:

1 Like