Counting lipid speckles/endosomes

Hi I’m a complete beginner to this program. I played CountSpeckles pipeline to orient myself but I can’t seem to find the right parameters.

I have fluorescently labeled cells (Red and Blue with puncta-like staining pattern at times) in culture. The dye is a membrane incorporating dye (DiI and DiD). There may be a membrane exchange/fragment updake phenotype that seems to change with certain experimental treatments. Ideally I’d like to quantify the following:

  1. % of blue cells having red speckles inside (and vice versa).
  2. Number of speckles in the cells.

Some main problems I don’t know how to get around:

  1. Non uniform puncta like stain makes it difficult to understand what color the cell was originally. By eye, you can tell blue or red cells mostly (cells with a uniform color tint or high population of a single color speckles). How to program the software to identify these correctly?

  2. (related to first concern above) How can the software distinguish the periphery of the cell in a crowded region of interest? Feels like software will have hard time to enumerate cells

Example image I’m trying to analyze is the attached.

Hi there,

I may have achieved some progress. Segmentation doesn’t look too right to me but it is a starting point. I edited few minor things in the original speckle pipeline.

Can you please let me know how I can improve the pipeline further? For the test run, I’m using “lossy” frame extracted from a compressed avi file (the experiment was a time course). Do you think that may contribute to problems during segmentation?

SpecklePipeline edited for DiI-DiD cells.cppipe (13.1 KB)

I think I fixed it by trial and error. Sharing my pipeline here in case someone wants to try.playgroundv3.cpproj (108.3 KB)