I am a bit confused by the integration tutorial. It indicates to use adata.raw to normalize and logarithmize the data from different samples I was just curious if this should be data that has already been normalized and logarithmized on an individual sample basis? The tutorial that mirrors seurat indicates to set the adata.raw parameter following normalization and running sc.pp.log1p so this would be what is present going into integration is that correct to do it this way? I know from Seurat they normally suggest running normalization on each individual sample prior to integration but I don’t believe they run a normalization on the merged sample set following individual normalization so I just wanted to check that I am doing things correctly. I appreciate any help you may be able to offer.