Speckle counting help

I am having trouble counting the foci of cells stained with cy3 and dap. I am trying to count the number of foci in the nuclei of each cell and the cytoplasm of each cell. I am trying to get a count of total specks per cell but also specifically just the nuclei and just the cytoplasm. I tried to base my pipeline off of the speckle example using similar modules: identify primary objects, enhance/ suppress features, mask image, identify primary objects, measure object intensity, measure object intensity, relate objects, export to spreadsheet.

The foci are not being quantified properly and I do not know what else I should add to the pipeline in order to get an accurate count. any help is appreciated! thanks

It’s hard to know what could be changed without more information (some sample images and a pipeline would be best, along with an explanation of “not counted properly”- too many, too few, broken down too much, lumped into groups, etc), but 3 major things to start-

  1. Have you looked at the output of the enhancement module to make sure it’s doing a good job of pulling out your particular foci? Play with the feature size to match it to your foci for best results.

  2. Assuming that the enhancement goes ok, and that your second “Identify primary objects” pipeline is actually picking up some of your foci, if you’re getting too many or too few you can try playing with the threshold detection method and correction factor; if it’s that it’s splitting them in half or grouping them together, adjust the settings for distinguishing and drawing the dividing lines between clumped objects.

  3. You say you want to measure the objects per nucleus, cytoplasm, and cell, but you don’t mention a cytoplasmic marker or any steps to divide the cells like that. Generally the pipeline would go something like “Identify Primary Objects” (nuclei) -> “Identify Secondary Objects” (cells, using the nuclei as seeds) -> "Identify Tertiary Objects (subtract the nucleus from the cell to get the cytoplasm). Then relate your foci to all 3 objects at the end. You’ll either need a cytoplasmic marker though, or to just set an arbitrary distance from the nucleus that you’ll call “cytoplasm”.

Hi Beth,
Thank you for your help, I am going to try and play with the threshold and correction factor now and see if I can get better results. In the mean time I have attached a screen shot of what happens when I use the current modules, and the .tiff files that I am trying to create this pipeline with. I appreciate any further insight you may have!
Thanks again,
speckle images.zip (2.9 MB)

Can’t tell for sure without knowing more about your experiment, but I think something got messed up in your channel names on export- your “dapi” looks like like speckles and your “cy3” is larger and looks more like a nucleus to me. Notice how speckly the image is you’re trying to draw nuclei on in your screenshot?

We are doing an in situ localization of proteins that are either cytoplasmic or nuclear localized. We are hoping to identify the individual specks that are in the nuclei or cytoplasm of the cells and then quantify them based off of their location. The nuclei is labeled with dapi and the speckles are labeled with cy3.

I switched the dapi and cy3 and you were right they were just labeled improperly! but I still need to refine the pipeline to get a more accurate reading.

1 Like