Dear all! When I use LIMMA to analyze microarray data, I use Array Weight parameter. I have some questions about it. The command is: #array weight arrayw <- arrayWeights(MAbe, design, weights=MAbe$weights) arrayweights <- targets for (m in 1:narrays){arrayweights$arrayw[m] <- arrayw[[m]]} write.table(arrayweights, file = file.path(resultdir, "arrayweights_BewteenArray.txt"),sep="\t", dec = ".",row.names= FALSE) # combine array and spot weights library(statmod) W <- matvec(MAbe$weights, arrayw) # Fit linear models fit <- lmFit(MAbe, design, method="ls", weights=W) # Compute contrast of interest cont.matrix <- makeContrasts (HvsL=L-H, levels=design) fit2 <- contrasts.fit(fit, cont.matrix) fit2 <- eBayes (fit2) # Toptable, remember to change name of generated files according to contrast toptable <- topTable(fit2,number=10000, adjust.method = "fdr") DE <- toptable$adj.P.Val[toptable$adj.P.Val<0.05] length(DE) # 2285 toptable <- topTable(fit2, number=2285, adjust.method = "fdr", sort.by="p") write.table(toptable, file = file.path( resultdir, "2285_between_ArrayWeight_fdr_CD8.txt"),sep="\t", dec = ".",row.names= FALSE) My questions are: (1) How is Log2(Fold Change) in the toptable gotten? If I don't use array weight, I can get log2(fold change) by (M(l)-M(h)). But when I use array weight, I don't know how to get it? (2) How to get the M value that is used in lmFit analysis? I know how to get the M value after normalization. I use this command. #loess normalize in arrays MAin <- normalizeWithinArrays(RG, method="loess", weights=RG$weights, bc.method="none") # write normalised data to TMEV-file - for clustering # Tab Delimited, Multiple Sample Files (TDMS) (*.txt) mMin <- MAin$M colnames(mMin) <- paste(colnames(mMin),targets$Pig,sep=".") clusterdata <- cbind(MAin$genes,mMin) FnameOutput <- file.path(resultdir,"Index_MAin_slides.txt") write.table(clusterdata, file=FnameOutput, sep="\t", row.names=FALSE) When I use array weight, I guess M value will be changed. Is it that? Thanks a lot for your attention! Best wishes, Yu GAO [[alternative HTML version deleted]]