Hello Everybody I want to identifiy differentially expressed genes from GSE23183 series file.But i am getting an error ( Error in ebayes(fit, ...) : No residual degrees of freedom in linear model fits ). Here is my R script:- *gse <- getGEO("GSE23183",GSEMatrix = TRUE) state <- phenoData(gse[[1]])$characteristics_ch1.3 state <- gsub("hiv disease status[:] soluble CD8[-]suppression","HivSolubleCD8",state) state <- gsub("hiv disease status[:] no soluble CD8[-]suppression","HivNoSolubleCD8",state) state <- gsub("hiv disease status[:] HIV suppression likely via direct contact [(]unrelated to CAF[)]","HivDirect",state) f <- factor(state) design <- model.matrix(~0+f) contrast.matrix <- makeContrasts(fHivDirect-fHivNoSolubleCD8,fHivDirect- fHivSolubleCD8,fHivNoSolubleCD8-fHivSolubleCD8,levels=design6) subtype <- gsub("\\.", "_", gsub("subtype: ", "",phenoData(gse[[1]])$characteristics_ch1.3)) pheno <- new("AnnotatedDataFrame", data = data.frame(subtype),varMetadata = data.frame(labelDescription = "subtype")) rownames(pheno@data) <- colnames(exprs(gse[[1]])) eset <- new("ExpressionSet", exprs = exprs(gse6[[1]]), phenoData = pheno) fit <- lmFit(eset,design) fit* * Error in ebayes(fit, ...) : No residual degrees of freedom in linear model fits* Can anybody help me out..Why i am getting this error.. With regards~ Reema Singh [[alternative HTML version deleted]]

On Tue, Jan 25, 2011 at 9:17 AM, Reema Singh <reema28sep@gmail.com> wrote: > Hello Everybody > > I want to identifiy differentially expressed genes from GSE23183 series > file.But i am getting an error ( Error in ebayes(fit, ...) : No residual > degrees of freedom in linear model fits ). Here is my R script:- > > > *gse <- getGEO("GSE23183",GSEMatrix = TRUE) > state <- phenoData(gse[[1]])$characteristics_ch1.3 > state <- gsub("hiv disease status[:] soluble > CD8[-]suppression","HivSolubleCD8",state) > state <- gsub("hiv disease status[:] no soluble > CD8[-]suppression","HivNoSolubleCD8",state) > state <- gsub("hiv disease status[:] HIV suppression likely via > direct contact [(]unrelated to CAF[)]","HivDirect",state) > f <- factor(state) > design <- model.matrix(~0+f) > contrast.matrix <- > > makeContrasts(fHivDirect-fHivNoSolubleCD8,fHivDirect- fHivSolubleCD8,fHivNoSolubleCD8-fHivSolubleCD8,levels=design6) > subtype <- gsub("\\.", "_", gsub("subtype: ", > "",phenoData(gse[[1]])$characteristics_ch1.3)) > pheno <- new("AnnotatedDataFrame", data = > data.frame(subtype),varMetadata = data.frame(labelDescription = "subtype")) > rownames(pheno@data) <- colnames(exprs(gse[[1]])) > eset <- new("ExpressionSet", exprs = exprs(gse6[[1]]), phenoData = > pheno) > fit <- lmFit(eset,design) > fit* > * Error in ebayes(fit, ...) : > No residual degrees of freedom in linear model fits* > > Can anybody help me out..Why i am getting this error.. > > Hi, Reema. Unfortunately, this GEO series has only three samples. You cannot perform statistical testing with just one sample per group. Therefore, the best you can do is to form ratios between the samples; that is what I would suggest you do and then rank genes based on fold change. Sean [[alternative HTML version deleted]]