Generating a cell-by-gene matrix

After pairwise alignment, the same coordinate system is shared between the spatial barcodes and the staining images.

However, analysis (e.g., clustering, pseudotime, DGE...) is performed on single cells, not on individual capture areas (0.6 μm in the current version of the protocol).

Here we show how to aggregate the barcoded spots-by-gene matrix into a cell-by-genes matrix, where cells are defined from the segmentation mask.

Segmentation of staining image

To create such a spatial cell-by-gene (\(M\times G\)) expression matrix, you will first need a segmentation mask.

We efficiently segment cells (or nuclei) from staining images using cellpose. For the H&E-stained tissue provided in this example, we used our fine-tuned model. Make sure to download it and save it into a new models folder that you need to create under the openst_e13_demo main folder.

You can run the segmentation on the previously created openst_demo_e13_mouse_spatial_beads_puck_collection_aligned.h5ad file, which contains the spatial transcriptome coordinates and staining image after coarse+fine pairwise alignment.

openst segment \
    --adata alignment/openst_demo_e13_mouse_spatial_beads_puck_collection_aligned.h5ad \
    --image-in 'uns/spatial_pairwise_aligned/staining_image_transformed' \
    --output-mask 'uns/spatial_pairwise_aligned/mask_transformed_0px' \
    --model models/HE_cellpose_rajewsky \
    --chunked \
    --gpu \
    --num-workers 8

After running this command, the segmentation mask is created and stored in the same --adata file, under the dataset uns/spatial_pairwise_aligned/mask_transformed_0px.

Assigning transcripts to segmented cells

Now, we aggregate the initial barcoded spots-by-gene matrix into a cells-by-genes matrix, by leveraging the segmentation mask.

Note

In this dataset, we did not specify --dilate-px. Since this model segments nuclei from H&E, we are referring to nuclei when we talk about cells. We use this terminology for consistency with the rest of the documentation.

This step allows you to aggregate capture spots by segmented cells:

openst transcript_assign \
    --adata alignment/openst_demo_e13_mouse_spatial_beads_puck_collection_aligned.h5ad \
    --spatial-key spatial_pairwise_aligned_fine \
    --mask-in-adata \
    --mask 'uns/spatial_pairwise_aligned/mask_transformed_0px' \
    --output alignment/openst_demo_e13_mouse_by_cell.h5ad

Expected output

That's it! After running the steps above, you will have that single alignment/openst_demo_e13_mouse_by_cell.h5ad file, which contains the transcriptomic information per segmented cell (nucleus).

Warning

Do not expect the files to be exactly the same, as there are several steps on the pipeline that are not deterministic (e.g., alignment, segmentation).

Now, the following sections provide some examples of exploratory data analysis that can be performed with this type of data.