Hi i have a proper count table of RNA-Seq and i have applied edgeR package for obtaining differentially expressed genes and I have obtained nice acceptable result. But now I am applying also DESeq over the same data but the pval and padj columns of the nbinomTest over them is strange, it is almost 1.00 or NA. Why is this so? thanks a lot for your answers in advance -- output of sessionInfo(): > head (res$padj) [1] 1 NA 1 1 1 1 -- Sent via the guest posting facility at bioconductor.org.

Hi On 2012-06-26 15:58, narges [guest] wrote: > i have a proper count table of RNA-Seq and i have applied edgeR package for obtaining differentially expressed genes and I have obtained nice acceptable result. > But now I am applying also DESeq over the same data but the pval and padj columns of the nbinomTest over them is strange, it is almost 1.00 or NA. > Why is this so? You know, we are not clairvoyants. Without any more information about your experiment and data, and without seeing the R code that you used to perform the analyses, any answer would just be guessing. Most likely you made a mistake in using the software. Or, maybe the two tools do come to different conclusions about the signal-to-noise ratio of your data. EdgeR and DESeq use similar but not identical methods. I understand that you prefer the result that says that your experiment worked, but this does not seem a good argument to decide which method is correct. Simon