Hi, SORRY, but I have a very basic question on dye swap. A set of cDNA microarray samples contains a few dye-swaps. For retrieving normalized, log2 transformed expression values (against a common reference for all samples), I wonder how I can make sure that the “dye-swap”-nature of the files is taken into account as I get the final spreadsheet of the expression values for all the Genepix files ! I have gone through limma which talks about fitting a linear model for each gene using design matrix. But, I wish to know just the expression values of the files (with dye-swaps taken care of in the final output). Please provide some suggestion(s)/pointer(s). Thanks a lot ! Cheers, Christina *The codes I used*: library (arrayQuality) targets <- readTargets("targetPMCI.txt") files <- c("3a.gpr","3b.gpr","3cDyeSwap.gpr","9bT.gpr","37a.gpr","37b.gpr","37c DyeSwap.gpr","61a.gpr", "61b.gpr","61cDyeSwap.gpr","75a.gpr","75bT.gpr","75cDyeSwap.gpr","76b. gpr","76cT.gpr","77a.gpr", "77bT.gpr","77c.gpr","78b.gpr","78cT.gpr","79a.gpr","79b.gpr","79cDYES WAP.gpr","80a.gpr", "80b.gpr","80cDyeSwap.gpr","81aT.gpr", "81b.gpr","81cDyeSwap.gpr", "82aT.gpr","82b.gpr","82cDyeSwap.gpr", "83a.gpr","83b.gpr","83cDyeSwap.gpr","84a.gpr","84b.gpr","84cDyeSwap.g pr","85a.gpr","85b.gpr", "85cDyeSwap.gpr","86a.gpr","86b.gpr", "86cDyeSwap.gpr","711aT.gpr") RG <- read.maimages(files,source="genepix") RG$printer <- getLayout(RG$genes) RG.b <- backgroundCorrect(RG, method="normexp", offset=50) library (convert) RG.Within <- normalizeWithinArrays (RG.b, method="loess") RG.Between <- normalizeBetweenArrays (RG.Within, method="Aquantile") genes <- RG.Between$genes log_ratios <- RG.Between$M Mean_LogIntensity <- RG.Between$A Log_Ratio_table <- cbind (genes,log_ratios) MeanLog_Intensity_table <- cbind (genes,Mean_LogIntensity) write.csv (Log_Ratio_table, file="Table of Log-ratio,M (cDNA, PMCI).csv") [[alternative HTML version deleted]]

Hi Christina, There's a number of ways to do that, all of which require you to first work out the direction on each array swap <- rep(1, length(files)) swap[grep("DyeSwap", files)] <- -1 swap Now swap contains +1 or -1. Now you can either fit a simple linear model: model <- model.matrix(~0+sweep) fit <- lmFit(RG.Between$M, model) Log_Ratio_DyeSwapped <- fit$coefficients -or- Log_Ratio_DyeSwapped <- sweep(RG.Between$M, 2, swap1, "*") hope that helps! Mark On 12/05/2009, at 6:02 PM, Christina Tigress wrote: > Hi, > > SORRY, but I have a very basic question on dye swap. > > A set of cDNA microarray samples contains a few dye-swaps. For > retrieving normalized, log2 transformed expression values (against a > common > reference for all samples), I wonder how I can make sure that the > ?dye-swap?-nature of the files is taken into account as I get the > final > spreadsheet of the expression values for all the Genepix files ! > > I have gone through limma which talks about fitting a linear model > for each > gene using design matrix. But, I wish to know just the expression > values of > the files (with dye-swaps taken care of in the final output). > > Please provide some suggestion(s)/pointer(s). > > Thanks a lot ! > > Cheers, > Christina > > > *The codes I used*: > > library (arrayQuality) > > targets <- readTargets("targetPMCI.txt") > > files <- > c > ("3a > .gpr > ","3b > .gpr > ","3cDyeSwap > .gpr","9bT.gpr","37a.gpr","37b.gpr","37cDyeSwap.gpr","61a.gpr", > > "61b > .gpr > ","61cDyeSwap > .gpr > ","75a > .gpr","75bT.gpr","75cDyeSwap.gpr","76b.gpr","76cT.gpr","77a.gpr", > > "77bT > .gpr > ","77c > .gpr > ","78b.gpr","78cT.gpr","79a.gpr","79b.gpr","79cDYESWAP.gpr","80a.gpr", > > "80b.gpr","80cDyeSwap.gpr","81aT.gpr", "81b.gpr","81cDyeSwap.gpr", > "82aT.gpr","82b.gpr","82cDyeSwap.gpr", > "83a > .gpr > ","83b > .gpr > ","83cDyeSwap > .gpr","84a.gpr","84b.gpr","84cDyeSwap.gpr","85a.gpr","85b.gpr", > > "85cDyeSwap.gpr","86a.gpr","86b.gpr", "86cDyeSwap.gpr","711aT.gpr") > > RG <- read.maimages(files,source="genepix") > > RG$printer <- getLayout(RG$genes) > > RG.b <- backgroundCorrect(RG, method="normexp", offset=50) > > library (convert) > > RG.Within <- normalizeWithinArrays (RG.b, method="loess") > > RG.Between <- normalizeBetweenArrays (RG.Within, method="Aquantile") > > genes <- RG.Between$genes > log_ratios <- RG.Between$M > Mean_LogIntensity <- RG.Between$A > > Log_Ratio_table <- cbind (genes,log_ratios) > MeanLog_Intensity_table <- cbind (genes,Mean_LogIntensity) > > write.csv (Log_Ratio_table, file="Table of Log-ratio,M (cDNA, > PMCI).csv") > > [[alternative HTML version deleted]] > > _______________________________________________ > Bioconductor mailing list > Bioconductor at stat.math.ethz.ch > https://stat.ethz.ch/mailman/listinfo/bioconductor > Search the archives: http://news.gmane.org/gmane.science.biology.informatics.conductor