normalization data transformation protocol
Data normalization was done on the raw .cel files for HG-U133A and HG-U133+2 using RPA [Lahti et al., 2011]. Custom array definition files were created using the customCDF R/Bioconductor, removing probes mapping to known SNPs, and summarizing probes for each gene with an ENSG identifier. Quality control was carried out using the R/Bioconductor package arrayQualityMetrics. The custom array definition (design) files have been submitted to ArrayExpress under accession numbers A-MEXP-2334 (for A-AFFY-33, HG-U133A) and A-MEXP-2334 (for A-AFFY-44, HG-U133+2).