I have a cohort of Affymetrix CEL file produced on 2 platforms that I would like to use them for a gene expression based classifier. I have removed artifactual bias between two batch using SVA package in R.
I am wondering if someone has experience scaling (Scale function in r) RMA Normalized Gene Expression data that has been used for gene expression based classification. Would it result in removing biological difference within each sample ?
I am also keen to know based on your experience in building expression based classifier whether working with log transformed data has better outcome ?
Thanks in advance.