Does log2 transformation normalize for sequencing depth or does it just log2 transform?

[s.l]

I’m trying to use the “Multiple expression tables” tool in order to complete a meta-analysis for RNA-seq datasets. I’m inputting raw gene counts from htseq-count (not normalized).

According to this paper (https://currentprotocols.onlinelibrary.wiley.com/doi/10.1002/cpz1.922), all tools under the normalization drop down menu should normalize your data for sequencing depth, and I just wanted to confirm it was the case for the “log2 transformation” option and that it was not just transforming fold changes into log2FC.

Thank you.

[jeff.xia]

Not sure if I fully understand your question,

• If your data is already normalized, choose “None”;
• If your data have only been adjusted for sequence depth (such as CPM), choose “Log2 Transformation” - Note LogFC is not a normalization method;
• If your data are raw count tables, use RLE or TMM to account for sequence depth and further normalization