Measuring area occupied within objects?

Hello,
I am trying to use CellProfiler to quantify mRNA expression within cells as part of the RNA Scope FISH assay. I’ve had success learning about the program from some of the published RNA Scope pipelines but have found a few caveats in my own experiments that have made this difficult, mainly that the transcripts we are measuring are very highly expressed within most cells and thus we can’t really discern individual puncta. This makes any count outputs unreliable as a fair number of “dots” identified are actually just a clump of brighter punctae. Optimally I would like to simplify the pipeline to measure the percent area of each cell that is occupied by each fluorophore by using thresholding, but I cannot find a way to measure area occupied within an object, as the MeasureAreaOccupied module only gives you the option to make binary measurements OR object measurements, but not both. I’ve been able to do something similar using ImageJ but I would really like to use CellProfiler since it seems much more reliable and efficient for managing large data sets. Is getting this type of data output possible in this program? I’ve attached an image below of what our raw data look like. Thanks!

I cannot help directly bust have you seen this

Thanks so much, I saw they had one a few weeks ago and didn’t know there was another - managed to secure a spot. :pray:

Hi @jannamoen,

CP could be used for RNA Scope analysis. If you could share the pipeline that you are trying with, we can help you better. Anyways to start with if you had not noticed,

& an earlier thread on the similar topic,

Regards,
Lakshmi
Fujifilm Wako Automation (Consultant)
www.wakoautomation.com
For CellProfiler training or optimised pipeline write to,
lakshmi.balasubramanian.contractor@fujifilm.com

Read more on our site.
Yokogawa CV8000 - The Ultimate in Confocal HCS
https://www.wakoautomation.com/products/yokogawa-high-content-imaging

Thanks Lakshmi, I actually have already utilized both of the resources you’ve linked, and the pipeline I’m using is essentially the same one found here (https://cellprofiler.org/examples/published_pipelines, Erben et al) with a few modifications (I’m only using 3 channels for my fluorescent assay and I only care about quantifying one). It works very well for identifying cells and associating the relevant fluorescent markers to cells, filtering, etc. The main problem is the output - this pipeline (and all others I have found) require that the data be punctate so that each puncta can be identified as a child object and thus counted within parent cells. My transcripts tend to be very heavily expressed and thus puncta cannot be individually identified. The best idea I have for getting around this would be to use thresholding and count the %coverage for each identified cell - in my case, the green channel is a cell type marker and the red channel is our transcript of interest. However I cannot seem to find a way to do this using CellProfiler. The only way it seems I can measure area occupied either outputs the area of each object (cells) or the binary area occupied across the entire image, not within individual cells.

Hi @jannamoen,

O.k. According to my understanding, you are interested in looking at how much the puncta (red) is been occupied in the Cells (Green). If this is right, there two ways I could think as follows,

  1. If you are interested in an image level, then you could use the segmented objects from both the channels & compute the Area & normalise them (i.e. Area of puncta/Area of Cells) where you could get the %.
  2. In case of object level, you should relate them as you had mentioned earlier (using “RelateObject” module) & now use them for similar normalisation (as mentioned in 1) individual cell level.
    For both of these you could use “CalculateMath” module in which you can do the image/objects using the measurements.

Hope this helps.

Regards,
Lakshmi
Fujifilm Wako Automation (Consultant)
www.wakoautomation.com
For CellProfiler training or optimised pipeline write to,
lakshmi.balasubramanian.contractor@fujifilm.com

Read more on our site.
Yokogawa CV8000 - The Ultimate in Confocal HCS
https://www.wakoautomation.com/products/yokogawa-high-content-imaging