I am trying to process the raw data in GSE20346. The processed data were normalized using a cubic spline algorithm (Illumina BeadStudio V2.0). However, I am trying to reproduce results from a paper published using GSE20346 where they normalized the data using normal-exponential background correction followed by quantile normalization. It does not look like control data were included in the publicly available data. I am using neqc (limma) to perform normexp background correction and quantile normalization, as my understanding is that it can approximate the background when no control data are available.
library(limma) example = read.ilmn("GSE20346_non-normalized.txt.gz") y = neqc(example)
The reason I don't believe this yields the correct results is because the paper published using GSE20346 released the normalized data. My results yield values ranging from 6-15, but the paper's processed values range from 0-14. Is there something in the workflow that I am missing, or am I misunderstanding how the neqc function works?
Maybe this isn't exactly what you're looking for but, but GSE20346 has been re-processed and analyzed on our Gemma database:
You can get differential expression analysis from the first link, either by disease state, timepoint or treatment.