Confusing code block in 4.image-and-segmentation-qc.py

[broadinstitute]

Confusing code block in 4.image-and-segmentation-qc.py

[ErinWeisbart]

Are you working off of CP151 data? Older batches don't have those measurements.

[ErinWeisbart]

We could add in for this step (as well as the PLLS and saturation plots) a logic that says if those columns can't be found then skip that unit in case a user accidentally removes those measurements from the CP pipeline?

[gwaybio]

Are you working off of CP151 data?

Yep

We could add in for this step (as well as the PLLS and saturation plots) a logic that says if those columns can't be found then skip that unit in case a user accidentally removes those measurements from the CP pipeline?

I am 100% for this strategy. In general, I'm a tad bit concerned that these QC scripts will fail out if CellProfiler output updates. This is exactly what I was trying to adapt to in the beginning! I will add the checks in #39

[ErinWeisbart]

Are you working off of CP151 data?

Yep

There are Correlation_Correlation_ columns in all the Image.csv files output by CP for all the CP151 data, so we need to take another look at the creation of the image_df?

[ErinWeisbart]

Oops! I take that back. They are there for CP151A2 and B2 but not for A1 and B1. So sorry about the confusion.

[ErinWeisbart]

They should be in all future batches, but adding in the check if column exists should solve the issue now and in the future :)

[gwaybio]

Got it! Super helpful comments

I add column checks in eff296e. I can also confirm that the code runs in the other wells when Correlation_Correlation_ is present