Dear all, this is to summarise recent changes to the "vsn" package that might be of interest: * In addition to normalising a set of arrays against each other, it is now possible to normalise arrays against an existing reference. This might be useful for applications to classification (when the reference is the training set) or in large studies when you want to add new arrays to an existing dataset without re-doing all normalisation. * Missing values are now allowed. * Glogs are now reported to base 2, in accordance with other packages. * Features on an array can be stratified according to e.g. GC-content, print-tip or cDNA-plate. * The function "vsn2" now returns an object of class "vsn" with the fit results; "justvsn" is a wrapper that returns the normalized and glog-transformed ExpressionSet. * The code has been cleaned up, is faster and should be more robust. The vignettes have been completely revised and are worth revisiting. See the vignettes of vsn >= 2.0.33 for details. You will need R 2.5 (currently: alpha) for this. Your feedback is welcome. Best wishes Wolfgang ------------------------------------------------------------------ Wolfgang Huber EBI/EMBL Cambridge UK http://www.ebi.ac.uk/huber