Hello, I would like to calculate Pearson correlation coefficient between the expression of two genes. The expression data comes from an experiment on a two channel custom Agilent array. The data is reported as normalized log10 ratio (Cy5/Cy3) representing test/reference. I tried to calculate the correlation on the ratio values but it seems to be wrong-( judging from the intersample correlation which is too low like 0.2 on average). The dataset is from GEO and there is also a file with test and reference intensities. I would appreciate any help, Regards, Maciej