Many thanks for the great talk earlier@Florian Wuennemann and @Erin Weisbart! I am very interested in building an nf-core module of the image feature extraction step out of cellprofiler. Would you mind showing some guidances of where to start?

Thanks! Glad you enjoyed the talk! The feature extraction step is what we call "analysis" and is in the nf-core-cellpainting here github.com/nf-core/cellpainting/blob/…/main.nf

thanks @Erin Weisbart for the reply! I believe this relies on a cellprofiler project file? like here:

path analysis_cppipe

Are you planning on building a pythonic way of running this step (some others in the future as well) with API calls?

Perhaps I misunderstood your original comment - by "out of cellprofiler" did you mean "using CellProfiler" or "outside of CellProfiler" ?

my bad. I wasn't clear. I meant more

outside of CellProfiler

. Like calling

regionprops

from scikit-image.

We have no immediate plans to bypass the use of CellProfiler, but as far as I understand it would be simple to support an alternative entrypoint in the workflow that swaps the analysis step for a non-CellProfiler version. FYI, CellProfiler v5 is on the horizon (no release date yet) and is separating all of the image processing into a cellprofiler-library python package so keeping the exact same feature extraction but doing it pythonically will possible "soon".

Thanks for the pointers! - Yes, I've attended once. But the slot keeps clashing with my shuttle back home... Will try my best for the next one 😀