Quality control and sample integration for Spatial Visium HD data

[ngochuyenpham]

Dear developers and users,

I am currently learning to analyze Visium HD data from the vignette that you provide https://satijalab.org/seurat/articles/visiumhd_analysis_vignette.

May I ask if you remove the bin that has low count or the one with high mitochondrial content? I noticed the violin plot in the vignette appears to have trimmed tails for certain cell types — is this because the data were preprocessed before being loaded into the vignette?

Should I perform a similar QC process as described for single-cell procedure? For example, remove low count and high mitochondrial bin. https://satijalab.org/seurat/articles/pbmc3k_tutorial.html
I also want to ask about the integration of Visium HD data. We have two mouse samples — one wild type and one with a genetic modification. I used the 8um level for analysis because it is an approximate single-cell resolution. Would you recommend integrating the two sample count matrices prior to downstream analysis? If so, which integration method would you suggest for Visium HD data?

Best regards,
Huyen