What type of normalization does removeBatchEffect function require as the input?

Question

[ev97]

I want to use removeBatchEffect from limma in my RNA data but I don't know how the input has to be. I know that I need log-expression values but I don't know what type of normalization is accepted.

removeBatchEffect(x, batch=NULL, batch2=NULL, covariates=NULL, design=matrix(1,ncol(x),1), ...)

According to the arguments...

x has to be numeric matrix, or any data object that can be processed by getEAWP containing log-expression values for a series of samples. Rows correspond to probes and columns to samples.

If we go to getEAWP we see that:

getEAWP(object)

0bject: any matrix-like object containing log-expression values. Can be an object of class MAList, EList, marrayNorm, PLMset, vsn, or any class inheriting from ExpressionSet, or any object that can be coerced to a numeric matrix.

Object should be a normalized data object. getEAWP will return an error if object is a non-normalized data object such as RGList or EListRaw, because these do not contain log-expression values.

It doesn't say anything about using a particular normalization.... So... Could it be possible to work with log2(TPM counts + 1)?

Thanks very much in advance

[Papyrus]

I would input depending on what methods you are using for the differential expression pipeline and dowsntream analyses? edgeR? limma-voom? DESeq2?, all these have different methods to provide values in the log scale e.g. logCPM(), cpm(), vst() or rlog()... (see for example this answer)

[ev97]

Thanks very much for your kind reply.

For downstream analyses I would like to use log2(TPM + 1), so I suppose that I could work directly with the result given by removeBatchEffect if I use the same log2 normalized method (log2(TPM + 1) as my input (?)

I just wanted to be sure, because as it didn't say (and I haven't seen anything) about the type of normalization... I was wondering if someone knows if it needs a particular normalization.