Batch effect removal (Combat)

Question

[GSAENZDEPIP]

Hi! I want to remove batch effect from a count data matrix (RNAseq experiment) using Combat function (sva package). I know that RemoveBatchEffect function (from Limma package) needs Log-transformed input, however Combat manual doesn't mention whether input data should be Log-transformed...

Thank you in advance!!

Combat manual: https://bioconductor.org/packages/release/bioc/manuals/sva/man/sva.pdf

[GSAENZDEPIP]

Thank you! I have just found the answer in the tutorial: "For sequencing data, which are often represented as counts, a more suitable model may involve the use of a moderated log function"

Answer 1

[h.mon]

The sva tutorial states one can extract the expression data from the expressionSet object:

The expression data can be obtained from the expression slot of the expression set.

edata = exprs(bladderEset)