I Freshly got the .CEL files from Affymetrix (Affymetrix GeneChip scanner) and was trying to analyse the data through dchip i have obtained .CDF files and the gene information files from affymetrix site for the proceedings. after loading the data in dchip, during the normalize and model step where it asks for baseline array[median probe intensities] by default it is considering the experimental data set as baseline (if it is purely based on median intensity i guess that should not be a problem or do i have to change it to the control data with comparatively good median intensities when compared to its replicates).