search

labsyspharm / mcmicro

Multiple-choice microscopy pipeline
https://mcmicro.org/
MIT License
110 stars 58 forks source link

Store channel names directly into .ome.tif #89

Open DenisSch opened 4 years ago

DenisSch commented 4 years ago

I noticed that some images get stopped earlier and they do not match the "marker.csv" anymore. @ArtemSokolov Suggestion: Would it be possible to count the amount of rcpnl files and double check with the length of the csv file? Should I create a separate python script or include this into nexflow?

ArtemSokolov commented 4 years ago

This can be done as a single nextflow line. Can you link a folder with .rcpnl files that I can use as an example for testing?

clarenceyapp commented 4 years ago

The number of rcpnl files alone may not be enough information for your intended application since rcpnl files can sometimes have 4 or 5 channels.

ArtemSokolov commented 4 years ago

Can throw a warning if number of lines in markers.csv is not a multiple of 4 or 5 relative to the number of .rcpnl files. Having it a warning won't prevent the pipeline from running, but would alert the user that there may be a discrepancy.

clarenceyapp commented 4 years ago

sorry, @ArtemSokolov . I was just using 4 and 5 as an example. It currently will be 4 or 5 which will be a helpful now, but as technology progresses, this number will increase. (aka someone else's problem)

ArtemSokolov commented 4 years ago

Can throw a warning if the number of lines in markers.csv and the number of rcpnl files are relatively prime? HAHA.

DenisSch commented 4 years ago

Let's do it at stage where it matters: Let's check the amount of layers in the tiff stack and compare it to the csv file?

DenisSch commented 4 years ago

Example: /n/scratch2/ds230/Z170/12

ArtemSokolov commented 4 years ago

If the check requires inspection of the file content (rather than the number of files), then I would say nextflow is not the appropriate place for it.

If you want to inspect a .tiff file, can we bake this check into the QC module that we talked about writing (e.g., before it starts comparing DNA channels across cycles)?

DenisSch commented 4 years ago

@jmuhlich Is there a way to include the names into the ome.tiff? That would pretty much solve all the issues with the mismatch of channels. Is this something on the horizon?