search

nasqar / deseq2shiny

a web-based R shiny application that wraps DESeq2 R package
http://nasqar.abudhabi.nyu.edu/deseq2shiny/
15 stars 7 forks source link

DESeq2 outputs "************" values #3

Open alexcrofts opened 3 years ago

alexcrofts commented 3 years ago

Hello,

First, great work on this app. I ran two different data sets (one mouse, one bacterial) and in both cases the DE analysis outputs some values for some genes as " ****". For example a log2foldchange value for one gene might be that long string of asterisks while all other values for the gene are expected number values. Then a different gene may have an adj p value of '****e-08' or while other values outputted for that gene are reasonable expected values. I tried to look this up in the DESeq2 manual to see if it is meaningful but it doesn't seem like this is standard output? These data are also not viewable within the shiny app and must be downloaded as .csv. I also tried viewing the .csv vs through the head command in linux to be sure it was a text viewer issue.

Any thoughts on preventing this or a way to see the actual value of these numbers? Thanks!

BrunoGuillotin commented 3 years ago

Hi, I tried for the first time this app and got the exact same issue. Many genes that seems to be DE have * always in one of the column. But the can be in every columns, baseMean log2FoldChange lfcSE stat padj , but never in the pvalue column.

Can the developers investigate on that ? should I consider this genes in my analysis ?