Hi, I was exploring batch effects in my data. At this stage, I don't want to introduce the 'variable of interest' (i.e. 'cancer' in the vignette/example given in sva.pdf that is used to get 'mod' - the model matrix). Should I use the mod0 matrix (i.e. the NULL matrix) as an input to the ComBat function (and get combat_edata2)? Will this correctly reflect the expression data corrected for batch effects? ###### Code ##### library(sva) library(bladderbatch) data(bladderdata) library(pamr) library(limma) pheno = pData(bladderEset) edata = exprs(bladderEset) mod = model.matrix(~as.factor(cancer), data=pheno) mod0 = model.matrix(~1,data=pheno) batch0 <- c(rep(1,12),rep(2,15),rep(3,30)) ## use mod combat_edata = ComBat(dat=edata, batch=batch, mod=mod, numCovs=NULL, par.prior=TRUE) ### use mod0 combat_edata2 = ComBat(dat=edata, batch=batch0, mod=mod0, numCovs=NULL, par.prior=TRUE) ################## thanks! [[alternative HTML version deleted]]