I am trying to use the DEseq package without replicates. When I look for significant hits all of them are False. Could this be a real result? Below are the commands that I enter. > countsTable <-read.delim("spore_readcounts.txt",header=TRUE,stringsAsFactors=TRUE) > rownames( countsTable ) <- countsTable$gene_id > countsTable <- countsTable[ , -1 ] > head(countsTable) > cds <- newCountDataSet( countsTable, conds ) > cds <- estimateSizeFactors( cds ) > sizeFactors( cds ) > head( counts( cds, normalized=TRUE ) ) > cds <- cds[ ,c( "1", "2" ) ] > cds <- estimateDispersions( cds, method="blind", sharingMode="fit- only" ) > res <- nbinomTest( cds, "1", "2" ) > plotDE( res ) Error: could not find function "plotDE" > addmargins( table( res_sig = res$padj < .1, res_sig = res$padj < .1 ) ) res_sig res_sig FALSE Sum FALSE 4960 4960 Sum 4960 4960 Many Thanks Melissa

Hi Melissa On 03/14/2012 03:50 PM, Melissa.Martin at lshtm.ac.uk wrote: > I am trying to use the DEseq package without replicates. When I look for significant hits all of them are False. Could this be a real result? Below are the commands that I enter. Yes, it is entirely plausible that this is a correct result. Remember that the "blind" mode is just a function to attempt to rescue a botched experimental design and get at least a few hits out of it. Whether this works or not depends on the nature of the data. As you did not tell us much about what kind of biology you are looking at, this is hard to say. Simon