[BioC] Normalize data across platforms

michael watson (IAH-C) michael.watson at bbsrc.ac.uk
Tue Mar 31 13:49:05 CEST 2009

Hi Steve

Quantile normalisation is a very conservative normalisation, and there are fears in some quarters that it may lead to over-normalisation.  However, as you're comparing between datasets, this may be what you're after.

Alternatively, have you considered using "housekeeping" or control genes?  These should be constant across arrays, experiments etc (if you believe in them) and so could provide a good normalisation factor.


-----Original Message-----
From: bioconductor-bounces at stat.math.ethz.ch on behalf of Steve Taylor
Sent: Tue 31/03/2009 12:44 PM
To: Bioconductor
Subject: [BioC] Normalize data across platforms

I have two sets of affy data CEL files. One set is from Hugene 1.0 ST arrays and the other from U133plus2 Arrays. I need to compare one set with another.

First I plan to use RMA to normalise the data set for each platform. I then plan to get a common reference id across the arrays, probably using ENSEMBL gene ID.
With the subset that have probes in common, what would be the best way to normalise across the arrays? Would quantile normalization using aroma.light be suitable?

Thanks for any advice,

Medical Sciences Division
Weatherall Institute of Molecular Medicine/Sir William Dunn School
Oxford University

Bioconductor mailing list
Bioconductor at stat.math.ethz.ch
Search the archives: http://news.gmane.org/gmane.science.biology.informatics.conductor

More information about the Bioconductor mailing list