sctransform on scRNA seq
Correct me if I'm wrong please :
does sctransform a method to eliminate batch effect and dropout phenomen in order to reduce zeros in the expression matrix ?
SCTransform is a way to normalize the counts, and batch effects at the manifold and clustering levels are usually taken care of via integration with Seurat integration, fastMNN, scVI, etc.
Imputation with e.g. MAGIC would be the method to "infer" gene counts for cells with dropout of certain genes. Use imputation with caution though as it's imperfect.
Traffic: 2534 users visited in the last hour