Working with NO Nuclear dye images


Dear Team,
For a long time, I’m struggling in figuring out, how to use CP for images where there is no nuclear dye used.
It has only one channel: GFP (green).
The images are basically from a time-lapse live cell imaging. However, I have saved each frame as a single tif file to be used in CP.
I have attached two of the 6 images here (jpg format).

So, is it possible to study these images?
If so, what pipeline we can use?
Below is the illumination correction pipeline that I developed.
SmoothOrEnhance (smoothing method: Enhance BrightRoundSpeckles (Tophat Filter))
RescaleIntensity (Rescaling method. (S) Stretch the image (0 to 1))

After this, when I run the actual pipeline for object identification and measurements, it couldn’t identify the primary objects, hence no result in the end.

PLz, help me in this issue.
Thanks for yr constant support.



I have a couple of questions so I can better understand what you are trying to do:

(1) What are your primary objects supposed to be: the cells or the nuclei? It looks like our want to make the cells as the primary object but I want to be sure…

(2) If I recall correctly, you wanted to study cytoplasm-nuclear translocation patterns, correct? If so, without a nuclear stain, this becomes a very difficult problem. If not, what are you trying to measure in this case?

Also, you have the module CorrectIllumination_Calculate in your pipeline but are not using CorrectIllumination_Apply to actually apply the results of CorrectIllumination_Calculate. Can you tell me what you hope to gain from using CorrectIllumination_Calculate?



Dear Mark,

My primary objects are nuclei so that I can monitor the cytoplasm to nuclear translocation in live cells.
Using a nuclear dye, translocation is getting inhibited in live cells (or some problem, we are trying to figure out). So, we are working with only one color (green), by which we can measure its intensity in both cytoplasm as well as nuclei. Without nuclear stain, its a challenge but is there any way, we can do it?

Is it possible to detect boundaries of nuclei and cytoplasm using one channel so that measurements can be made up and ratios can be examined?

Using CorrectIllumination_Calculate in the pipeline, I want to do illumination correction which I’m applying in next pipeline where object identification and measurements are being calculated?

Plz, let me know and thanks for your help.
Mridul KK



A significant problem you’re going to deal with if you have no nuclear dye is how you’re going to determine where the nuclear border is when the florescence intensity is roughly equal in both the cytoplasm and the nucleus. In this case, you will not be able to see the boundary between the two. If a human observer cannot see the border, CellProfiler will not be able to either.

Perhaps you can detect the nuclear border at a point when the florescence is largely restricted to the cytoplasm only (I assume this occurs at the early time points of your experiment?). In this case, you might be able to distinguish the two compartments and save the nuclear location as a file. This file can be used at later time points as a mask to create the two compartments in the later time points when the florescence blurs the difference in the one channel.

However, this approach will only work if the cells in general (and the nucleus specifically) are motionless between the early and late time points; the position of the mask nuclei must be constant. Is this true in your case? If so, then a pipeline might be more straightforward to develop which does what is needed.



Hi Mark,

Thanks for the reply with suggestions.
I do understand the challenge of boundary detection and we are working on this issue of second channel.
However, in the meantime, I was still working with the hope to develop “some” method of working with CP.

Will let you if I come-up with some breakthrough.

Mridul KK