DEseq results -- foldchange is 0/inf

Question

[Nan]

So after feature counts of RNA-seq bam file, I have an count file. I input the count file into DEseq, and got results which contain foldchange values such as 0/inf/NA, so how can I deal with these values when I want to use foldchange to filter out most up-/down- regulated genes?

So for foldchange == NA, I think this case can directly dropped. But what about foldchange ==0/inf?

Thank you.

id          baseMean  baseMeanA       baseMeanB  foldChange  log2FoldChange  pval      padj
SOCS4       1834      2321            1348       0.580       -0.7844         0.00038   0.844
NPIPA3      34.1155   68.23175774754  0          0           Inf             7.51E-09  7.71E-05
AL627309.5  2.0225    0               4.045      Inf         Inf             0.434     1
AP002833.2  0         0               0          NA          NA              NA        NA

[Ram]

Please use the formatting bar (especially the code option) to present your post better. You can use backticks for inline code (`text` becomes text), or select a chunk of text and use the highlighted button to format it as a code block. I've done it for you this time.

P.S.: You can also pretty print tabular text using column, as shown here.

Answer 1

[Chirag Parsania]

Hi,

Inf and NA occurred due to 0 in one of the sample (baseMeanA or baseMeanB). The common practice to deal with this problem is to add small number (e.g. 0.1 ) to normalised expression value (e.g basemean/FPKM/RPKM) and recalculate the fold change.

[Nan]

Yes, this is a way to avoid 0/inf values. But I was worried if it is biologically reasonable?

[Chirag Parsania]

It should be ok as you are adding small constant value to all the genes. Therefore, it nullifies any possibility of bias.