jump-cellpainting / datasets

Images and other data from the JUMP Cell Painting Consortium
BSD 3-Clause "New" or "Revised" License
149 stars 13 forks source link

Exact commit hashes for Cellprofiler pipelines used in JUMP production #82

Closed hanslovsky closed 9 months ago

hanslovsky commented 9 months ago

Is there documentation for the exact version and commit hash of the cellprofiler pipelines that were used by each source for analysis? https://github.com/broadinstitute/imaging-platform-pipelines/tree/master/JUMP_production#production-pipelines is linked in the paper but there are two pipeline (JUMP_analysis_v2.cppipe and JUMP_analysis_v3.cppipe). Did all sources use the same pipeline (v3?) at the same commit hash?

Thank you

niranjchandrasekaran commented 9 months ago

I believe everyone used v3. Tagging @bethac07, who can confirm.

bethac07 commented 9 months ago

Everyone used the measurements in v3, but classical segmentation is typically not robust enough to not need tweaking across multiple sites - sometimes not even across multiple batches. So segmentation steps were not held constant across sites - for sources that sent Experiment.csv files, those will record exact settings for each plate, otherwise you'd need to get pipelines from each source and confirm which batch(es) each was used for.

hanslovsky commented 9 months ago

Great, thank you @bethac07 and @niranjchandrasekaran!