Mapping of Xenium 5k data

[shanggutianzhen]

Hello, after updating squidpy, it seems to be able to read fluorescent negatives, but there is no entry to read HE negatives (after align). And the drawing speed seems to be very slow. My computer is configured with i9 13900-kf,4090 and 128G memory, but it takes about 30-40 minutes to complete the output of a picture when making a picture

[timtreis]

How large is the picture itself? When rendering with matplotlib, large pictures can take quite a bit. You could try to convert our data to SpatialData and then use spatialdata-plot for rendering - there we have several performance improvements for large images.

[shanggutianzhen]

Wow，Thanks for your reply.It's about 8GB in our HE.But in the Xenium data,HE need to be corrected to apply the data,how to do it in the squidpy?

[timtreis]

Hey @shanggutianzhen, I don't fully get your question, sorry. You could try reading your data with our Xenium reader (https://spatialdata.scverse.org/projects/io/en/latest/generated/spatialdata_io.xenium.html) directly into SpatialData though

[shanggutianzhen]

Oh，i am sorry to borther you,but in our data,HE slices were obtained by using other scanning machines alone, so they were not aligned (HE images and fluorescence signals).

[timtreis]

So with "corrected" you mean "aligned"? There are mostly automatic tools like https://github.com/JEFworks-Lab/STalign or you can do it semi-manually as f.e. described here: https://spatialdata.scverse.org/en/stable/tutorials/notebooks/notebooks/examples/alignment_using_landmarks.html