Dear Jack, lmFit() allows NA values in the expression values but not in the design matrix. If you must put NA values in a column of the design, then you need to remove them like this: anyna <- applyis.na(miRNA),1,any) fit <- lmFit(mRNA[,!anyna], miRNA[!anyna,]) This code identifies rows of the design matrix containing NAs, then removes those samples from the data. lmFit does not do this automatically, because I feel that this is something that a user should make a decision about deliberately. Best wishes Gordon > Date: Fri, 22 Jun 2012 09:59:28 -0400 > From: Yao Chen <chenyao.bioinfor at="" gmail.com=""> > To: bioconductor at r-project.org > Subject: [BioC] [Limma] Calculate the relation between mRNA and miRNA > > Hi all, > > I have a problem to use lmFit calculated the correlation between mRNA and > miRNA, because my miRNA data contained "NA" values. > > So if I use :lmFit(mRNA,miRNA), I got the message "Error in qr.default(x) : > NA/NaN/Inf in foreign function call (arg 1)". > > The original "lm" function allow to included "NA", why lmFit can't? > > Thanks, > > Jack ______________________________________________________________________ The information in this email is confidential and intend...{{dropped:4}}