Backsub Changes All Fluorescence to DAPI

[eisele24]

Hi there! I am running into an issue after running background subtraction. After registration with Ashlar and image cropping, my resultant ome.tif images are well aligned and each channel has distinct staining. However, when I run background subtraction, the resulting ome.tif file looks like DAPI in every channel of the hyperstack. I also noticed that the registered ome.tif input also changes to have DAPI staining in every channel after running background subtraction. Have you encountered this before? I am attaching my markers.csv and a screenshot of my params.yml file for review, in case there is something that is causing this. Thanks so much!

markers.csv

[ArtemSokolov]

Tagging @kbestak who developed that module.

[kbestak]

Hi @eisele24, thank you for bringing this up as I have not seen this behaviour before.

To help with troubleshooting, could you please send the contents of the .command.out file accessed in the work directory?

You can find the path to the work directory from the run output similar to this:

executor >  slurm (1)
[-        ] illumination                -
[-        ] registration:ashlar         -
[27/d9d75b] background:backsub (1)      | 1 of 1 ✔
[-        ] dearray:coreograph          -
[-        ] dearray:roadie:runTask      -
[-        ] segmentation:roadie:runTask -
[-        ] segmentation:worker         -
[-        ] segmentation:s3seg          -

The work directory in this case would be work/27/d9d75b...

Additionally, I'm puzzled that the input also changes. Would it be possible to rerun the step outside of MCMICRO to see whether the behaviour persists? You can see the documentation here: https://github.com/SchapiroLabor/Background_subtraction/

Thanks, I hope we understand the underlying issue soon!

[eisele24]

Hi kbestak, thank you for your reply and apologies for just getting back to you! Here are the contents of my .command.out file:

Pixel size overwrite not specified Channel DNA_1 (0) processed, no background subtraction Channel CD34 (1) processed, no background subtraction Channel Granzyme B (2) processed, background subtraction Channel CD56 (3) processed, background subtraction Channel DNA_2 (4) processed, no background subtraction Channel CD8a (5) processed, no background subtraction Channel CD11c (6) processed, background subtraction Channel ICAM1 (7) processed, background subtraction Channel DNA_3 (8) processed, no background subtraction Channel CD20 (9) processed, no background subtraction Channel CD127 (10) processed, background subtraction Channel CD117 (11) processed, background subtraction Channel DNA_4 (12) processed, no background subtraction Channel CD45 (13) processed, no background subtraction Channel CD31 (14) processed, background subtraction Channel CD4 (15) processed, background subtraction Channel DNA_5 (16) processed, no background subtraction Channel Ki67 (17) processed, no background subtraction Channel CD69 (18) processed, background subtraction Channel Collagen_type1 (19) processed, background subtraction Channel DNA_6 (20) processed, no background subtraction Channel CD15 (21) processed, no background subtraction Channel CD163 (22) processed, background subtraction Channel NKp46 (23) processed, background subtraction Channel DNA_7 (24) processed, no background subtraction Channel CD68 (25) processed, no background subtraction Channel CD3 (26) processed, background subtraction Channel CD14 (27) processed, background subtraction Channel DNA_8 (28) processed, no background subtraction Channel SMA (29) processed, no background subtraction Channel GrzmK (30) processed, background subtraction Channel CD49a (31) processed, background subtraction Channel DNA_9 (32) processed, no background subtraction Channel CD94 (33) processed, no background subtraction Channel NKp80 (34) processed, background subtraction Channel CD10 (35) processed, background subtraction Channel DNA_10 (36) processed, no background subtraction Channel NKG2A (37) processed, no background subtraction Channel NKG2D (38) processed, background subtraction Channel EMPTY_4 (39) processed, background subtraction Channel DNA_11 (40) processed, no background subtraction Channel PD-1 (41) processed, no background subtraction Channel CD103 (42) processed, background subtraction Channel PAX8 (43) processed, background subtraction Channel DNA_12 (44) processed, no background subtraction Channel CD5 (45) processed, no background subtraction Channel CD79a (46) processed, background subtraction Channel FOXP3 (47) processed, background subtraction Channel DNA_13 (48) processed, no background subtraction Channel Pan_Keratin (49) processed, no background subtraction Channel IL-15 (50) processed, background subtraction Channel EMPTY_3 (51) processed, background subtraction Channel DNA_14 (52) processed, no background subtraction Channel SMA (53) processed, no background subtraction Channel AF568 (54) processed, no background subtraction Channel AF647 (55) processed, no background subtraction [(1024, 1024), (1024, 1024), (1024, 1024), (1024, 1024)] Script finished in 8m 34s

I also am puzzled about the input file changing. I used this representative patient (CC077) to rerun the entire pipeline (illumination, ashlar stitching and registration, and background) and checked the ome.tif files as I went to ensure the issue wasn't arising sooner. I can't quite figure out what to make of it! Let me know if there is anything else I can send over, I'll also try to run outside of McMicro in the meantime!

[eisele24]

hi there @kbestak ! I was just wondering if the contents of the command.out file provided any clarity of what could be going on?

[kbestak]

Hi @eisele24, thanks for the ping!
The output seems fully expected and the script prints as if working normally. Were you by any chance able to run the tool outside of MCMICRO to check whether the behaviour repeats?

[eisele24]

Great, @kbestak that's relieving to hear! Admittedly, I'm not too familiar with running this pipeline so I had hesitations about trying to run the tool outside of MCMICRO. Do you mind helping walk through what that process looks like?

[kbestak]

Here is the tool's repo with documentation: https://github.com/SchapiroLabor/Background_subtraction
In brief, you would give the registered image as input, as well as the same markers.csv as you did already while using the exact same image that was cached when you ran MCMICRO. Could you open an issue there and refer to here?

For further context here, could you please write the full command you used in addition to the params and markers file you've already provided? Thanks!