Render_points super pixelized after subsetting SpatialData

Help spatialdata
1

Hello,
I cropped my SpatialData object using coordinates (load the coordinates csv with spatialdata_io.xenium_explorer_selection(file) and subset using polygon_query() ).
I then plot the expression of some transcripts on the cropped_sdata with:

cropped_sdata.pl.render_points(
        "transcripts", #SpatialElement of Points to plot
        color="feature_name", 
        groups=["draxin","eomesa"], #entries of feature_name
        palette=["orange", "blue"],
        size=0.8
    ).pl.show(title=f"{brain}")

The result is super pixelized and I wonder if that is expected. Here is the image:

image
image563×619 62.1 KB

Note that when I plot the shapes, the result is normal:

render_shapes
render_shapes373×484 90.7 KB

Thank you for any insight!

2

Hi, it looks like the rasterization with datashader is being used automatically for the cropped data. Could you please report in spatialdata-plot? Meanwhile passing method=’matplotlib’ should be a workaround.

3

Should be fixed by Fix datashader resolution collapse on cropped coordinate offsets by timtreis · Pull Request #669 · scverse/spatialdata-plot · GitHub

2 Likes