Morphology of cancer cell line

Hi, I’m researching a brain cancer cell line and want to know how they are affected by some particular biological substance regarding cell morphology. I have taken so many pictures under fluorescence microscope after treatment with the special substance and ethanol as a control. I now want to use cellprofiler to be able to detect if there is any difference in the appearance of the cells treated with this substance compared to the cells treated with ethanol. I use hoechst (Dapi) to stain the nucleus and phalloidin (Alexa) to stain the actin filament. I’m interested in changes in cell morphology that I can see in the actin filament.
I do not know how cellprofiler can help me in this case to find the cell forms I am looking for and if there really is any difference between the cell morphology of the cells treated with the substance and the cells treated with ethanol. I also do not know which pipeline to use in cellprofiler. Can you explain and tell me how cellprofiler can help me in this case?

Segment the cells then extract shape features with the module MeasureObjectSizeShape. Sometines it can be useful to also include skeleton features (module MeasureObjectSkeleton). From there, if you can annotate examples of different morphology classes, you can take a supervised learning approach to classify all cells into the known morphology classes. If no morphology classes are known a priori then take an exploratory approach using clustering.

Thank you, what do you mean by segment cells?

Segmentation is the process of finding objects in an image (here cells). This is usually done by assigning labels to each pixel and grouping pixels with the same label defines one object. In CellProfiler, this is performed by the IdentifyObjects modules (i.e. IdentifyPrimaryObjects, IdentifySecondaryObjects…). A common approach is to start by identifying/segmenting nuclei with IdentifyPrimaryObjects then identify the cell boundaries in the actin channel with IdentifySecondaryObjects.
If you need more detailed pointers, please post an example image.

6 M.tif (4.1 MB)

Here I attach a picture of my cells. You see in the middle of the picture some long and elongated cells and I first want to use cell profiles for the images to be analyzed and segmented as you said and cell profiles can distinguish between the nuclei which are colored in blue and the actin filaments which are colored in red. Then I want to train the cell profiler analyst program to look for the elongated cells and quantify them in both treated cells and untreated cells so that I get an idea if the actual number of cells that are elongated is more in the treated cells than in the untreated cells (because I have seen such long cells in untreated cells too, but much less than I have seen in treated cells) then I can conclude that my substance has affected the morphology of the cells and that the appearance of the cells has changed after treatment with the particular substance. I need to first use cell profiler to create a property file and then use that property file in cell profiles analyst to learn and train the program to look for the cells I am interested in. And it is also important for me to know how to create the property file correctly. Thanks for your help in advance.

Another problem that exists is that when I run the IdentifyPrimaryObjects module, an error message appears that says that my image should be gray, but the image I have is red and blue. What is the problem do you think?

Regarding IdentifyPrimaryObjects, this module is designed to operate on a single channel. You’ll probably want to run colour images through the ColorToGrey module first to split apart the different wavelengths.