Discriminate between transfected and untransfected cells


#1

I am new at using CellProfiler. Let me start by saying that I think it is pretty easy to work with and has a nice interface and informative help-functions. Unfortunately I am running into some novice problems.

I am trying to quantitate the Nuclear/Cytoplasmic distribution of a GFP-tagged protein, which I transfect into cells. So some cells will be GFP-positive, whereas others are GFP-negative.
I would like to filter out the GFP-negative cells, but so far I have been unable to come up with the right solution.

The pipeline files I have composed from scratch is at follows:
Pixel Size: 1

Pipeline:
LoadImages
IdentifyPrimAutomatic
SubtractBackground
IdentifySecondary
SubtractBackground
IdentifyTertiarySubregion
MeasureObjectIntensity
CalculateRatios
CalculateRatios
ExportToExcel

Could you let me know of a solution to only define cytoplasm in the GFP-positive cells? I have tried some filter and exclude modules, but so far I’ve been unable to come up with anything that works.

A confocal image illustrating the GFP and nuclear signals is posted at:

home.hetnet.nl/~rvamn/vector_ch00.tif
home.hetnet.nl/~rvamn/vector_ch01.tif

and screen dumps that hopefully illustrate the problem are found at:

home.hetnet.nl/~rvamn/identify_nuclei.fig
home.hetnet.nl/~rvamn/identify_cells.fig
home.hetnet.nl/~rvamn/identify_cytoplasm.fig

Thanks a lot,
Renee


#2

Hi Renee,

Please read he help for the module FilterByObjectMeasurement. With this module, you can filter objects that fit a specific measurement requirement. In your case, I think you can identify the nuclei, then measure the GFP intensity within the nuclei and filter based on this measurement. The Integrated Intensity of GFP within the nuclei will most likely be the best measurement to filter by.

Good luck and please ask for help if you have any problems.

Regards,
Mike


#3

Dear Mike,

Thanks for the reply. I will give the filter option another try. I did try it before, but I was having a problem with the filter settings because not all GFP postive cells have a GFP-signal in the nucleus. I’ll try again though.


#4

Hi Renee,

You can take the same approach using the cell outline instead of the nucleus. You might also be able to get a better nucleus segmentation by increasing the smoothing of the image which is an option in IdentifyPrimAutomatic.

Mike


#5

Hi Mike,

I got it to work, thanks for the suggestions. I just had some trouble figuring out at what point the FilterbyObjectMeasurement had to go (I now have it before the IdentifyTertiary), and putting in the values to use was a bit trial and error. But it’s working now.
Thanks!