Using CellProfiler with CellProfiler Analyst to Identify and Count Cells

Sample image and/or macro code

Snap-1195_c1.tif (5.1 MB) Snap-1195_c2.tif (5.1 MB)
RBPMS_DAPI.cppipe (26.0 KB)

Background

Hi everyone,
I’m trying to use CellProfiler Analyst to count the number of cells in a series of images. I have 2 channels for each image (DAPI and Alexa). My cells tend to be found in clusters and I’m trying to use DAPI as a seed to separate the cells labeled with Alexa.

Analysis goals

So far, I’ve had pretty good success using just a CellProfiler pipeline, but I think CPA could improve my accuracy. I want to be able to process my images in CellProfiler and then use CPA to further refine identification and count the total number of Alexa and DAPI positive cells.

Challenges

I’ve tried to be thorough in CellProfiler, but I can’t CPA to load anything other than blank image segments (no cells on them). I’d appreciate any help - thanks in advance!

Hi @Philippe_D_Onofrio,

First of all, are you saying that you think you can use CPA to improve your segmentation? Maybe I misunderstand but I’m not how you could achieve that in CPA.

Re your issue, I think it might be due to your input images being RGB. Do you know why this is? I wouldn’t expect these single channel images to both be separate RGB images and to CPA each of your “channels” has 3 channels as you can see from that error in the console as seen in the screenshot below.

Hi @lmurphy,
thanks for your reply. I’m not really trying to use CPA to improve segmentation, instead I was trying to use it after segmentation with CP to improve identification of my cells. Do you think this is the best way to go about this? My goal is to count only the cells that are double-labeled with DAPI and Alexa, and to make the pipeline work on a very large data set where there can be some variations in lighting and quality of labeling (although nothing too drastic). I\m not sure why the images are in RGB, but I will re-upload them as single-channels.

Oh, that sounds fine to be using CPA for. I just misunderstood your first post I guess and was concerned you though it might somehow be able to be used for segmentation.

So your images aren’t RGB when you are using them in CellProfiler?

No problem - thanks for your thoroughness! About my images: when I put them into CellProfiler, they are RGB, but I put them into Grayscale using the NamesAndTypes module.RBPMS_DAPI.cppipe (26.0 KB)