Dear BioC What is currently considered the best method of normalisation where the assumption that most genes are unchanged is invalid? Is it best to apply the Li & Wong algorithm or are there other methods that I should consider? I wish to normalise and analyse the following: 1. A focussed oligonucleotide array. Although this included a good number invariant "control" genes, most of the genes on the arrays are expected to change. 2. Arrays in which methylating agents are studied, again a large number genes on the array are expected to change. 3. Taqman RT-PCR cards. I would be grateful if you could tell me if I need to consider different methods for: A. oligonucleotide single channel focused arrays, B. dual channel spotted focussed arrays Thanks a million for your advice in this. I appreciate your help. Best wishes Aedin