Registration creates lines/squares on image

[Lucas-Maciel]

Hi,

I have RNAScope WSI from 3 cycles that I would like to register. Cycle 1 and 2 looked very nice, but in cycle 3 some black lines started to appear. I've tried to use both cycle 1 or 2 as reference. but this doesn't change it.

This first screenshot is DAPI cycle 1 + 2.

Here is just the DAPI signal from cycle 3, where we can see very thin black lines.

This is the first time I use the tool, but this issue was also encountered by the person who recommended me this tool.

Thanks for your attention
Lucas

[Yu-AnChen]

Apologies for the slow response. I'm happy to take a look if you could share your images. I can also be reached via email yu-an_chen [at] hms.harvard.edu

[Lucas-Maciel]

I just sent an email with my images. Thank you

[Yu-AnChen]

Hi @Lucas-Maciel, thanks for being patient and sharing your images! I think I've identified and fixed the bug. Please get the latest version of palom (v2023.8.2) and lmk if the issue persists.

Here's a jupyter notebook for documenting my process. Interestingly VGP_cycle3_ROI1 is rotated 180° relative to the other two images. If this is common, the detect_flip_rotate=True in as in In [7]: should handle image flipping. Note that in In [11], I was only selecting the DNA channel (reader.pyramid[0][0]) to write to the output file.