Creating a Pipeline in CellProfiler for Automated Analysis of Punctate Staining in the Nucleus of 2D Confocal Images

Sample image

Positive Control:

Negative Control:

Background

Hello, I am an undergraduate student working this summer with my mentor to streamline the analysis (quantification) of NFATs inside the nucleus of mouse lung epithelial cells. Images are obtained using confocal microscopy, then are transferred to Fiji where we enhance the contrast of the channels and compile into one image. The green shows the lung epithelial cells (club cells), blue DAPI stain for the nuclei, and the red punctate staining shows the NFAT proteins.

Analysis goals

Our goal is to use CellProfiler to analyze the amount of NFAT inside the nuclei to streamline and enhance our image analysis process focusing on quantification of punctate staining in the nucleus.

Challenges

My mentor and I are both unfamiliar with the CellProfiler program, and I am hoping for some assistance in the process of creating a pipeline. We believe we will need one similar to the example pipeline, “Speckle Counting” and will need to use ImageMath (mentioned in forum post below) to subtract out the green club cell in order to focus on NFATs located in the nucleus only.

From our research thus far, we believe our pipeline will resemble the following:

  • Import/convert .nd2 files

  • uniform image adjustment (i.e. contrast)

  • define ROI (i.e. nucleus)

  • quantify pixel density in ROI

  • export graph and tables

As an undergraduate, I have little experience with this type of work and I am unfamiliar with this program. I do however, have interest in research and am excited to build upon my knowledge by learning a new program! Any and all advice, assistance, or resources you may be able to provide would be very helpful and appreciated. Thank you for your time!

https://forum.image.sc/t/pipeline-for-counting-cells-with-only-membrane-staining-in-cellprofiler/32447/2

Hi @MadelynR,

Welcome to the image.sc forum!

The problem you describe can certainly be approached with CellProfiler. I’d recommend the following:

  • This video tutorial is excellent for an introduction to principles of image analysis and understanding how CellProfiler works: CellProfiler Workshop - YouTube
  • I think this article is an excellent way to think about image analysis: When To Say ‘Good Enough’ | Carpenter Lab
  • After watching the video, try to build your own CellProfiler pipeline. I think your general steps are correct – a rough outline is that you will import the .nd2 files, then segment the nuclei, then measure the amount of NFAT signal within the nucleus (probably using the MeasureObjectIntensity module), then export the data with ExportDataToSpreadsheet
  • Note that when you run your analysis, you’ll want to do so on the original confocal microscopy image files rather than the compiled .jpg images (jpeg results in the loss of data as the image is compressed)
  • If you run into any challenges when building the pipeline, you can reach back out on the forum with an example pipeline and images and we’ll try to help.

Good luck and congrats on undertaking this research project!
Pearl