SCTransform normalised data or regular log normalised data for assessing gene-gene correlations with pySCENIC?

[lucygarner]

Hi,

I would like to use pySCENIC (https://www.nature.com/articles/s41596-020-0336-2) to identify modules of genes predicted to be regulated by a given transcription factor (regulons). I was wondering whether scRNA-seq data that has been normalised using SCTransform is suitable for this approach or whether this type of normalisation interferes with gene-gene correlation structures? The alternative options would be to use raw count data or count data normalised using the NormalizeData() function in Seurat.

Thanks for any insights that you can give.

Best wishes,
Lucy