I need to measure the nuclear translocation of the protein of interest in my cells. So I need to calculate the ratio of the nuclear to the cytoplasm protein content. I downloaded the SBS pipeline then I added color to gray module to it because my images are colored. I also added 2 correct illumination modules as well as a correct illumination apply module. I went through the settings of each module however, it seems that there is something that I cannot adjust properly. Attached please find three images, each image has 2 channels Dapi and the protein of interest in red.
I really tried so hard to get the pipeline to work on my own but it seems that I cannot.
I would really appreciate your help with it, please .
Thank you very much.
translocation-p.cp (15.1 KB)
You did pretty well, though I made a couple crucial changes (imho), so take a look at the attached pipeline. Here are the main changes:
- CorrectIlluminationCalculate: Your images do not need much (if at all) illumination correction. And if you are doing per image correction (i.e. not per plate) then I suggest the “Background” method, paired with the “Subtract” option in CorrectIlluminationApply. I made these changes and you can see that they are not really affecting the raw images at all. Also, your block size MUST be large enough to include some
- IdentifyPrimaryObjects: The nuclear segmentation needed some improvement, and so I changed the “declumping” from Intensity to Shape/Shape. I tweaked a couple other parameters but Shape helped the most.
- IdentifySecondaryObjects: With translocation assays, if there is a chance that you won’t have ANY staining in the cytoplasm, then it is safest to use the “Distance-N” option so that the region expands regardless of the protein marker. But if you think this is not the case, then change back to DIstance-B. I expanded the region a bit wider to 10 pixels too, but that is your choice.
- I removed a space in a named output in CalculateMath. Try to use alphanumeric and underscore characters only, as it is safer.
translocation-p_DLogan.cp (14.8 KB)
Thank you very much. This actually terrific. I really appreciate your help!
I ran the pipeline and I think it works perfectly. However with some images where the cells are spindly I do not think that I can identify the cytoplasm properly. Attached please find an example where the image has both types of cell; the rounded cells as well as the spindly ones. Please let me know what you think.
One more thing please, as I understand the histogram shows the y-axis (the number of the cells) and the x-axis(the math ratio). However, the final histogram that I get does not include all the cells in the images I have uploaded. It only represents the cells in the second image- as cycle 2- for example. So how can I adjust it to include all the cells in the 2 or three images I upload for the run.
Unfortunately, the Display modules do not permit the cumulative display of a given measurement.
If you want a true segmentation based on the cytoplasm marker (keeping note of the translocation assay caveat that I pointed out previously), then change IdentifySecondaryObjects method to “Propagation”. Adjust the Threshold Correction Factor as you see fit. I tried with the new image and it looks reasonable at the default parameters to me.
As for the Histogram for all images, try running your experiment and then try Data Tools -> DisplayHistogram and load in your output file. I think it will display all the data (though I haven’t confirmed this). But yes, as Mark says, definitely modules within the pipeline don’t have access to all the images you have loaded.