Dear limma users and all bioconductor's users, I have four questions about the subject. I analyzed our data in the following way: library(limma) library(affy) todos <- read.matrix("todos.txt",sep="\t") todos <- log2(todos) todos.norm <- normalize.quantiles(todos) design <- model.matrix(~-1+factor ...... cor <- duplicateCorrelation(...... todos.fit <- lmFit(todos.norm,design,....... namegene <- read.table("nameGENE.txt"...... uniquegenename <- uniquegenelist(namegene,...... contrast.matrix <- makeContrasts(T-C,.... todos.fit2 <- contrasts.fit(todos.fit,..... todos.fit3 <- eBayes(todos.fit2) todos.fit3$genes <- uniquegenename TC <- topTable(todos.fit3,number=100,coef=1,adjust="fdr") write.table(TC,"TC.txt",sep="\t") First question: this script is OK? Second question: Although already to have propagated here in the list some messages in respect of this subject, none of them gave me certainty. Therefore, I go to ask again, because I still have doubts. Fold Change is the ratio between Treated (T) and Control (C). Correct? Third question: In the majority of the cases, the data are transformed into log in base 2. Then, in these cases, I have: (Log2 T / Log2 C). Correct? In this case, Fold Change would be equal to 2^(Log2 T / Log2 C). OK? Fourth question: Exists some case where the ratio between the T and C are made before the Log transformation? For example: Log2(T/C). In the way that I made our analyses, value "M", in topTable, would be the described one in the third question? How it was gotten? First was gotten average among repetitions of the treatment and average among repetitions of the control and, then, obtained the ratio? I ask because I made some empirical attempts and I did not obtain how the "M" value was obtained. If this is very complicated, please, you are free to skip it because I would not understand if it is very complicated. My excuses for the long e-mail. Thanks very much Best wishes Marcelo