Do I need adding 1 pseudocount (prevent dividing by 0)?

[wmseq]

Hi every one,
I am wandering if I need adding 1 or 0.1 pseudocount to prevent dividing by 0 for edgeR analysis, after absent genes were removed from my count file. Hope to get your advice.
Thanks a lot!
Richard

[dpryan]

I guess it depends on what you're doing that having an infinite ratio is problematic.

[wmseq]

Devon,
glad to meet you again, thanks a lot!
As you know, after the acount table for rnaseq analysis is created, the absent genes, ambigous genes,.. should be remove from the table. It seems that another step, adding 1 or 0.1 to the zero of a gene count in a sample, is needed. I don't know its purpose, so send this question on the furm.

[dpryan]

Ah yeah, it's pretty rare that you really need to do that. As a general rule, don't muck with the counts table unless you really need to.