Thanks for your answer. I ultimately built my own CDF.
I need to normalize data from the Affymetrix Mouse Gene 1.1 ST array to produce expression calls by exon. I currently get gene-level calls using oligo:
library(oligo)
celFiles <- list.celfiles(DIRLOC, full.names = TRUE)
affyExpressionFS <- read.celfiles(celFiles, pkgname = "pd.mogene.1.1.st.v1")
ppData <- rma(affyExpressionFS)
I know I can get probeset-level calls by using the target parameter of oligo's rma(). However, there are usually several probesets per exon on this chip, and I would rather not have to reconstitute this for all genes (e.g., map probesets to exons and create new meta-probesets that are actually exons). When Speed's group did their Affymetrix chip comparison they did exactly this for human data. Has anyone already solved this problem for mouse?
1 answer
You should be able to do this with the Aroma package...
You can obtain a custom CDF file for the MoGene 1.1 ST array and then perform exon-by-exon summarization.
Log in to answer this question.