Exported intensity values are below set threshhold value

Image Analysis
#1

Hello ,

I have been trying to measure fluorescence intensity of cells stained for IL-10 (FITC_green) and TNF-alpha (Rhodamine_red) which are both in the cytoplasm. I use the nucleus (DAPI_blue), as primary objects to locate the secondary objects (Cells for TNF signal and Cells 2 for IL-10 signal). For the identification of the secondary object, I manually put in a thresh-hold value, 0.005. However after running my anaylsis, I found out the CP also picked up fluorescence intensities lower than the set threshhold value.
I expected all the recognised intensities to be above or equal to my threshhold, which is 0.005… Please I need to understand what is going on?

Here is a sheet containing the exported intensities
040717_m1gm_images.xlsx (88.1 KB)

This is the pipeline used for the analysis
trial pipeline 2_integrated.cppipe (23.8 KB)

Here is an image set

Thanks for your help.
Didi.

#2

Hello Didi,
Some image processing steps such as smoothing may cause a change in intensity. Introduction of a mask and/or adjustment of parameters such as “regularization factor”, not filling holes in identified objects and not shrinking smaller objects prior to subtraction can all help minimize such effect. Also, the cytoplasm would be best to look at which will require to identify tertiary objects. Please see the file below and let us know if you have any questions. The median and mean intensity of the cytoplasm are equal to or greater than 0.005 after adjustment of all mentioned parameters except mask introduction.

trial pipeline 2_integratedVC.cpproj (678.4 KB)

Kind regards,
VC

#3

Thanks for you suggestions and help VC! I really appreciate. I will re-analyse my images including the identify tertiary object module.

Thanks again,
Didi.

#4

Hello,

I have received a lot of help from this forum for which I am very grateful.
However, I still need a bit of help to really understanding how CP measures
fluorescence intensity. I have attached my question (which is quite
lengthy, please bear with me), alongside the pipeline used for my analysis
and also the images that were analysed.

Thanks so much for your help.

Didi

green_intensity.cppipe (18.4 KB)

#5

Hi,

Your images and pipeline showed up, but it doesn’t seem like your question did; please update and let us know the question!

#6

Hello,

Please find attached a word document titled ‘*question’. *I started with
explaining how I went about my analysis. My question is on the 4th page of
the document highlighted in yellow.

Thanks,
Didi.

#7

Ahh, that’s the problem- the forum doesn’t allow word documents for security reasons, and even if we did it’d be better for you to put the text directly here so that your question is searchable for future users. I’m slightly concerned by the fact that you’re saying it’s four pages long though- perhaps you can summarize it somewhat?

#8

Oh! I see…

In trying to figure out how the measuring intensity module works on the cell profiler, I identified my secondary object (a surface marker) using 4 different manually set thresholds, 0.005, 0.01, 0.1 and 1. I subsequently measured the secondary object size and intensity.
With 0.005 I got the following measurements:

image.png742×634 342 KB

I went further to increase the threshold to 0.01 and got the following measurements:

image.png737×629 338 KB

I noticed the mean cell area reduced which makes sense because object areas with intensities less than 0.01 was cut off. This resulted in the mean intensity being higher.

With 0.1, I got these:

image.png738×634 331 KB

I noticed the mean cell area became same as the nuclei area. This made me think the threshold was higher than the highest intensity on the image. However looking at the intensity measurements, a mean of 0.012 (lower than the mean calculated when 0.01 was used as the threshold) was generated. Given that the threshold is high, I didn’t expect CP to give me any values.

I went on to use 1 as my threshold value and I got these measurements:

image.png741×633 333 KB

I got exactly same measurements generated using 0.1 as threshold.

Questions: Please can you explain why I got similar values using thresholds 0.1 and 1? Also, does CP put the set object identification threshold into consideration while calculating object intensity?

Thanks,
Didi.

#9

I think your main confusion is stemming from what a secondary object is- it’s the primary object (in your case the nucleus) PLUS the area around it, not merely the area around it alone. That’s why when you increased your threshold to 0.1 or 1(higher than any intensity value in your cells channel), the nucleus and the cells are now the same size- the “cells” objects and the “nuclei” objects are now for all intents and purposes the same. If you just want the area that’s outside the nuclei, you should do as @vchernys suggested and create a TertiaryObject, which will be just the cytoplasmic area.

I think you’ll figure out the answer to the rest of your questions based on that understanding.

#10

Hello,

I finally understand the concept secondary and tertiary objects; and truly,
I figured the answers to the rest of my questions.
I really appreciate the assistance! :slight_smile:

Thanks,
Didi