> From: Matt Wayland <mwayland@hgmp.mrc.ac.uk> > To: Bioconductor@stat.math.ethz.ch > Subject: [BioC] Affy RNA digestion plots > I have found the RNA digestion plots useful, but have > been disappointed that no scale is provided on the ordinate axis (mean > intensity). To try to get round this problem I printed the means.by.number > data (see below) and plotted these values against position in probeset using > MS Excel. I was surprised to find that the graphs produced in MS Excel had > quite different profiles to those displayed in the RNA digestion plots. What is plotted is the same standardized data used to produce slopes. First overall chip mean is subtraced chipwise from all means.by.number. Then each of the means.by.number are divided by corresponding standard errors. Finally the values 1 to number of chips are added chipwise to the plotted values, to stagger the lines and make them easier to see. There is no scale, because after this it would not be helpful. means.by.number and standard errors are made available partly so you can look at raw data. Future releases will include other plotting options.

Hi! What does the p-value of the RNA degradation test actually say and where do things begin to become critical? For instance, I have data here where the p-values are rather low (hgu133A chips) and in fact the slopes are positive! What is the appropriate interpretation? No RNA degradation? Something funny?? > summaryAffyRNAdeg(deg) X1 X2 K1 K2 L1 L2 slope 1.44 1.57 1.28 1.24 2.15 1.51 pvalue 5.91e-08 2.74e-08 7.11e-07 6.65e-07 8.93e-09 2.88e-08 Cheers, Hinnerk