Question about multiple spatial samples

[atrostleSW]

Hello,

Thank you for this helpful software!
I have some visium data with biological replicates across several sequencing slides, what is the best way to make use of all data in a co-expression analysis?

Best,
Alex

[smorabit]

Hi Alex,

Thanks for your interest in hdWGCNA. This is an open ended question so I am not sure that I can provide meaningful advice without knowing more details. For example, do you have timepoints, treatments, conditions, disease groups, genotypes, etc? What tissue? Do you have a specific niche / cell type / tissue region of interest?

[atrostleSW]

Hi Sam, thanks for responding! I am working with mouse digits. They are quite small, and to get enough power, we put several per slide. We are interested in how the different tissue types signal to each other, so you can imagine for example a little bit of skin on each replicate sitting next to its own little bit of connective tissue.

Best,
Alex

[smorabit]

Thanks for the additional info! I think that it should be feasible to run hdWGCNA on the entire dataset. To start, I recommend following this tutorial.

Also, as noted in other GitHub issues (#278), some key changes in Seurat 5.1 have broken the compatibility between Seurat and hdWGCNA for Visium datasets, so at this time we recommend using Seurat 5.0 or using Seurat 4 for this analysis.

[atrostleSW]

Thanks for the clarification!
Do you have any advice for replicates across multiple sequencing slides? I was not sure if I should try to run your tool on all at once with transforms to the imagerow/imagecol information or if I should run each slides and then follow your tutorial for Consensus network analysis