Segmented CODEX data in QuPath csv to squidpy

[andreevg04]

Dear Community,

I have stitched CODEX images directly after the run and I segemented the image in QuPath. The result is a csv where each row is a cell with multiple columns of Marker Expressions, X and Y Centroids and other information such as in which compartment the cell is etc.
I want to start analyze the neighborhoods in squidpy or to start putting the file in spatialdata-io but I do fail generating the needed anndata object for it to run in squidpy.
Could you help me with some kind of support.
*I dont get the raw images or fcs file from the experiment. Only ome.tiffs that I segment etc. in QuPath.
Best wishes!!!!

[LucaMarconato]

HI @andreevg04, thanks for your interest in our packages.

I would suggest to start using the codex() reader from spatialdata-io (https://github.com/scverse/spatialdata-io/blob/main/src/spatialdata_io/readers/codex.py), and then to enrich the object (or in alternative to create a new SpatialData object), adding a new table and new segmentation from the results of QuPath.

Please you can find here:

• how to manipulate/enrich spatialdata objects
• define new elements (e.g. a new table and new segmentation mask
• more details on how to work with tables
• discussion about squidpy integration

I hope this help!