Hi Group, I have a doubt in regard to cyclic loess method. normalize.loess(data,epsilon=1,log.it=F,span=0.4,maxit=4). As I understand that the number of iterations (maxit) should be performed until there is no convergence. 1) Would it make a difference if I log and do the cyclic loess or use unlogged dataset to perform cyclic loess(by mentioning in the function log.it=F)? I am asking this because in the microarray dataset of 48 samples when I log then it goes on and on and never gives "no convergence" for atleast 10 iterations that I tested.But when I use unlogged dataset it gives no convergence after 4 iterations. Could anyone tell me if by not performing the log, can I still get a good normalized dataset by cyclic loess? Waiting for your reply, Thanks, Viritha [[alternative HTML version deleted]]