Cell segmentation leveraging membrane-specific staining

[farshadf]

Hi Bin2cell team,

I was wondering if there is any possibility of using membrane-specific staining to improve the cell segmentation, preferably in IF settings. This would be very helpful for dense areas or where there is a considerable overlap between cells.

I could not specifically find anything related to that in the documentation or code of bin2cell or stardist. If I have missed it, I would appreciate you referring me the section.

Best,
Farshad

[nadavyayon]

Hi,

bin2cell only has a stardist model for segmentation and that is only a nuclear model. So membrane stainings are not part of bin2cell. You can segment your cells with other tools. As long as you save these segmentations in the same way we currently do you should be able use these segmentations for the rest of the pipeline.

N

[ktpolanski]

To the best of my knowledge, 10X doesn't currently support membrane staining in Visium HD, and I'm pretty sure that when they will the software will leverage the segmentation like they do for Xenium. That said, to reiterate #11 and what Nadav just said, bin2cell is designed to be modular and you can easily replace the segmentation with a mask made via another tool. Keep us posted if you find a good one - would be useful to know.