GFP Uptake Count Per cell

Hi There,

I would be grateful if I could get some help on the output part of the cellProfiler.
I am working with MCF-7 cells that are up uptaking GFP.

I believe I have managed to make the pipeline:

-Identifyprimaryobjects (Nuc)
-identifySecondaryObjects (Nuc again, masking around the nuclei)
-IdentifySecondaryObjects (Cell mask/Membrane)
-MeasureObjectsizeshape (Nuc)
-FilterObjects (Nuc)
-IdentifyPrimaryObjects (GFP)

  • EnhanceOrSupressFeatures (GFP)
    -RelateobjectsIntensity (GFP in cell)
  • MeasureobjectIntensity (Green GFP in cell)
    -Graytocolour
    -Overlayoutlines
    -ExportTospreadsheet
  1. Does my pipeline look ok?
  2. I basically want to know how much GFP is inside each cell, how can I see that in the excel file?

I would be most grateful!

Kind regards,

Hi @Raza,

Welcome to the image.sc forum!

We’re happy to help you troubleshoot and evaluate this image analysis workflow. Your described pipeline sounds like it should have the measurement data that you’re hoping for (GFP instead of cells). When you run the pipeline the ExportToSpreadsheet module should save a table for each object that you create. The “Cells.csv” table should contain a column for the IntegratedIntensity of GFP (which is measured by the MeasureObjectIntensity module). You’re looking for a column named something like “Intensity_IntegratedIntensity_GFP” (names will vary depending on the names you assign in your pipeline).

I hope that helps. If you still have questions, a good next step would be to upload your project and an example image, and we can take a look at the pipeline directly. Good luck!
Pearl

Hi Pearl,

Thank you so much for your quick response, I am not sure I am able to see in the “Cell.csv” File.

I will attach my Pipeline and the images for you to have a look at if that is possible. Thank you!

Also, wanted to know if it gives me the GFP count per cell in that image file. For example, if there is 18 cells detected, I wanted to know how much GFP is taken up in each cell, is there a way I can see that?

These images are from a Z-stack, but I am just trying to figure out the basics before I go and analyse 3 planes.

Thank you so much again!

GFP Pipeline 2021.cpproj (1.2 MB) 210319_SP_EV_3HRS_x60_1_GFP_z1_Ch1-T2.tif (171.2 KB) 210319_SP_EV_3HRS_x60_1_GFP_z1_Ch2-T1.tif (812.6 KB) 210319_SP_EV_3HRS_x60_1_GFP_z1_ChS1-T1.tif (579.8 KB)

Hi @Raza,

A few responses:

The data files are saved out as .csv files. These are similar to a spreadsheet and can be opened in text editors, spreadsheet programs like Excel or Numbers, or imported into a Google spreadsheet. There are also various free online CSV viewers if you search “CSV viewer”.

When I run your pipeline and open the “MyExpt_Cells.csv” file, there is a column named “Children_GFP_Count” that has a count of the number of GFP objects per each cell.

A few notes about your pipeline:

  • your nuclear segmentation thresholding looks great, well done!
  • you may want to choose to “discard objects outside the diameter range” and set different boundaries for the sizes in order to eliminate very small dots from being included as nuclei
  • your image for detecting cytoplasm is challenging to use to detect cells because the staining is very dark in the interior of the cytoplasm. One thing that can help to detect dim cells like this is to either use the “Log transform before thresholding?” option or to use the ImageMath module to raise the original image by a factor of 0.5. Doing so can help increase the pixel intensity of the dimmest values:
  • Finally, at least to my eyes, your IdentifyPrimaryObjects module for detecting GFP appears to be detecting a lot of very dim objects as true signal. I’d recommend trying out the RobustBackground method for this segmentation, since you have a very dark background with a few dispersed bright spots:

This tutorial covers a lot of what I discussed above regarding optimizing segmentation settings and the underlying theory. Highly recommended! CellProfiler Workshop - YouTube

Cheers,
Pearl