CosMx Strategy

1. Export data

Export the data from the AtoMx platform, which should include:

  • flatFiles

  • RawFiles (contains Morphology2D and Misc)

Note

In AtoMx v1.4, all RawFiles are included; in later versions, Spot files are excluded.

2. Unzip files

Run the following command from within flatFiles to unzip all the .csv.gz files:

gunzip *.csv.gz

3. Add raw images

Use the bash script CellLabels.sh (see more in Tools).

  • Update the SOURCE_DIR and DEST_DIR variables in the script

  • Make the script executable and run:

chmod +x CellLabels.sh
./CellLabels.sh

This will create a CellLabels folder inside flatFiles.

4. Add CellComposite folder

From the Morphology2D directory, choose either:

  • Composite .jpg images (CellComposite)

  • Raw multichannel .TIF images (≈200× larger)

Both are located in RawFiles/CellStatsDir.

If CellComposite is absent or unsatisfactory, generate it with:

python tools/make_composite_revised_image.py

This creates multiple folders inside Morphology2D. The most relevant are composite and composite_autocontrast

Rename your chosen folder to CellComposite and move it into the flatFiles folder. You can also use the renaming_composite.sh script from tools/ to standardize image names.

5. Finalize data

Once flatFiles contains both CellComposite/Morphology2D and CellLabels, you can:

  • Import and create the .zarr object:

    python src/qc/CosMx_QC.py

  • Create the Napari visualization (see How to use Napari)

  • Perform QC using the same code, including defining sample FOVs