I use CellProfiler to quantitate individual cell sizes in tissues. I am thankful to you guys, since CP works best so far among softwares that I have tested for single cell quantifications. However, still have some false positive data points (due to overlapped cells, false particles from fragmented cells etc).

I thought CPA can help me to efficiently sort out false data points. I really want to try it out (I even played with CPA examples), however, learning CPA isn’t as intuitive as CP for me. I want to see scatter plots or data distributions of all data points to locate outliers. Honestly I don’t know how to start with. I have already generated data spreadsheet with CP. Ideally, I want to see scatter or histogram plots of my data set next to original images with labeled identified particles. This would help me to determine outliers with great confidence.

As an example, I attach my csv data, pipeline and images. Thanks.

Tg2576_2_30x2_FilteredNuclei.csv (77.3 KB)
Tg2576_2_30x2_cytoplasm.csv (78.7 KB)
Tg2576_2_30x2.mat (210 KB)

This is one of the things CPA does best, and we will soon be releasing a bunch of new tools that make it even more useful for doing quality control.

My suggestion to you is to use the ExportToDatabase feature in CP 2.0 and have it generate a CPA properties file and an SQLite database for you. When you run CPA, you need only open the properties file and you should be able to go from there.

One thing you need to note is that you’ll need to make sure you choose to output a “single object table” in ExportToDatabase, This way you’re nuclei and cytoplasm data will be condensed into the same per-object table because CPA requires exactly 1 per-image table and 1 per-object table.

CPA is also distributed with a properties_README.txt file and a manual which you should take a look at. They should help you get an idea of what you need to get going.