Abstract
Four different normalisation techniques were applied for the corrections of fluorescence data generated by a cDNA microarray experiment. Correction for inaccurate signals and possible bias induced by fluorescence intensity, background intensity and dye effect were used in different combinations. Results of the present study highlight a pronounced role for the normalisation techniques in the absolute number of genes different expressed and a low concordance between different methods. Moreover, a significant effect of the dependent variable used, mean or median fluorescence intensity, was observed.
Key words: