Cell counter and overlapping quantification

Hello all,

I need some help for my experiment.
I have to analyse neuronal markers (GAD2 and GAD67). I did a containing with this 2 markers (cytoplasmic markers) and now I would like to know how much cell expressed GAD2, how much expressed GAD67 and how much expressed both. I attached a picture of the 2 staining. (on this picture there are not many but in the other yes that is why I can’t count it manually)

Could you help me in this way ?

Best regards,


Hello Gabrielle. In my opinion there is no feasible way to automate analysis of these images. The DAPI stain is workable, but these GAD stains are not only very dim, but the background is huge. I have no experience with GAD antibodies as neuronal markers… why don’t you use MAP2? Do you have a lot of these, and are they all like this?

Edit: Ok, I read now that you want to analyse neuronal markers. So you want to know how well they work? And these are proven to be neuron specific? Do you have a control marker? (Like map2)

Here is an image with a rainbow LUT, you can see more cleary how sparse the signal is. If this was my experiment I would count only a few of these signals as cells…

Hello Sverre,

Thank you for your answer.
I use different neuronal markers because I want to see if there is a difference in the neurone populations in my experiment compare to the control. GAD2/67 are GABAergic neurones and I test Glutaminase, CamKII as well.
I agree there is a lot of background…
This is a picture around hippocampus where I have less cells but in the cortex I have really a lot and it is more complicate.
I do not have a control marker. I think GAD2/67 are neurone specific…

Thank you for your image. Is it a merge of the 3 channels?

Thanks a lot !

No, this is just your 2nd image with a LUT that colors by intensity, where red is the brightest and blue the dimmest. It is purely for vizualisation. (Image >> Lookup Tables >> Rainbow (RGB), in FIJI)

Can you share one of these cortical images? And yes GAD2/67 are neuronspecific.

Yes but staining is really not the best … :frowning:

I agree, I would be wary of using this for quantification… Do you have a ton of these already or can you change your staining protocol?

Quite a lot I would say … I will do it manually I think