Needing help to quantifying fluorescent intensity

Hi to every one,

I’m trying to quantify fluorescent intensity in my images to quantify YAP expression in these cells. My problem is if I quantify the total intensity of YAP (red channel) per cell, the mean is diluted by all the black present in the signal inside the cell because YAP is not expressed equally in all the compartments. So, I tried to make a binary mask after a gaussian blurr (1.5px) and a median filter (1 px) and thresholding to catch all the signal from 20 to 255. After this, I make several operations of erode and open trying to have a good area and delete some points produced in the thresholding step.

From this binary image, I created a ROI that I used to delimit the quantification in the original YAP IF image, keeping mainly with the perinuclear area which is my area of interest.

Do you guys think it is a good way to quantify YAP or is it another way to do this? I have attached some photos of the process.

Many thanks.

@VictorVitio

Ideally - you don’t want to threshold the same channel being measured for intensity.

Take a look at this older post by @oburri that explains why:

As you can see in your snapshots - by setting a threshold of YAP - you are cutting off lower intensity value areas… so you are biasing the results off-the-bat.

So - take a step back… what is your scientific question here? What are you hoping to achieve with these measurements? Perhaps knowing that - we can better help you define a way to answer it. Try to stay away for science-specific jargon…

Too - sticking with whole-cell measures is better if you can (but pay attention to cell-cell overlaps in this case - can you easily distinguish one cell from another?) - maybe the mean is not an appropriate measure in this case… look at the histogram. Perhaps median is more reflective? Others may have ideas here…

2 Likes

Have you attempted the same technique with an inverted image?

Thank for your reply. I was thinking about your answer. My biological question here is if YAP has a differential expression in a mutant cell compared with WT in order to characterize the if the mutant has the same activation capability as the WT.

I think other option is doing the same way but only cutting the black pixels in the image to have the real area where the YAP signal is. after this, I can divide the intensity by the number of cells in the image to estimate the expression per cell. What do you think?

Thank for your reply. I cannot understand the sense of work in the inverted image. If I do that, I expect to have a measure of the region in which it doesn’t have a strong signal of YAP. I think it is not really informative. What do you expect if I do in the way you said?

Many thanks

Hello Victor, You know what you are looking for so if you invert the image it may be easier to segment, then re-invert the image after segmentation for the original image. It sometimes works very well.

I do understand your concern with the suggestion, is does sound strange however I use the technique quite often when the objects I am working are at the lower end of the intensity histogram.

Good Luck, If I can be of further assistance, just let me know.

Bob

@VictorVitio

Ideally - you use a different marker/channel to delineate your cells. Imagine you have a cell that has very low expression… so lots of ‘zero’ pixels within - but then you ignore them and it falsely inflates the intensity you measure within. So again - I would advise against this method…

2 Likes