generate_celltype_data with raw or SCTransformed data

[artemisiat]

Thank you for this comprehensive package. I was wondering whether you recommend using raw or other assay (sct/integrated) counts as input for the construction of the CellTypeData file. Thanks!

[Al-Murphy]

Hey, great to hear you are finding the package useful! We advise using normalized gene count matrix with EWCE, this is described in our extended vignette documentation: https://nathanskene.github.io/EWCE/articles/extended.html#create-a-celltypedataset. Here we discuss using scTransform to normalise for differences due to cell size, then linearly scale. We have functionality for this:

cortex_mrna$exp_scT_normed <- EWCE::sct_normalize(cortex_mrna$exp) 

[artemisiat]

Thanks! Does this function work any differently than the Seurat one? I wouldn't expect so.

[Al-Murphy]

I'm not sure what the differences are but other normalisation approaches should also be appropriate

[artemisiat]

thanks!