sum 1 to whole matrix counts in differential expression analysis. its ok?

[fgajardoe]

Hello everybody.

In our laboratory we are analyzing RNAseq data to do a differential expression analysis. Now we are using DESeq and we want to select the most differentially expressed genes using just fold change value. The RNAseq experiment has no replicates, so i dont care about statistics. it is just a descriptive study.

The problem is that we cannot sort the list because of "Inf" and "-Inf" fold change values. We are thinking to sum 1 to the whole count matrix to avoid the zero counts problem. Is this a good idea? Is there some inconvenients?

Thanks in advance

[dpryan]

You might have a little better luck with GFold, or with simply setting a minimum read threshold in both samples.