Dear all, I have time course data (processed and have expression values as log2 fold changes) with and without drug at various time points. I want to do a clustering experiment with data. I read from various papers that before doing clustering, the data is normalized using mean, median or centroid mean. I have 2 questions: 1. Why do I have to do normalization using mean or median or centroid mean. what is going to happen if I standardize. what is specifically centroid mean and how it differs from mean and median. 2. How can I do it in R. are there any packages /functions available. thank you very much. ps ________________________________________________________________ ____________________ Never miss a thing. Make Yahoo your home page.