RMA background correction normalization only

Question

[akutasame]

Hi,

Does the RMA method for background correction and normalization (not summarization) give log2 transformed intensities? or which step gives log transformed values?

Thank you in advance, Good luck in your studies

[akutasame]

I really need an answer, please make a comment if you have an idea.

[Jason Chen]

Are you using the expresso function in the "affy" Bioconductor package? I believe in that one, the log2 transformation is done as part of the RMA summarization step (also in medianpolish), and not in the previous steps. But, you could always check what the output looks like to make sure.

[akutasame]

Nope, I am doing probe level Anova, I did RMA background correction and quantile normalization only, I did not do summarization. But I just found that RMA background correction is taking log2 base raw intensities. http://www.plexdb.org/modules/documentation/RMAexplained.pdf

Thank you for your comment

Answer 1

[Kevin Blighe]

The RMA method by definition produces log [base 2] (log2) intensities. The steps are:

1. background correction
2. quantile normalisation
3. log2 transformation

Summarisation at the probe or probe-set level can be controlled via the target parameter that is passed to the rma() function. Depending on the array type, a probe-set may encompass a single gene or a single exon. Try it out yourself.

Kevin