Off topic:Choosing appropriate FPKM cutoff for finding differential expressed genes

Question

[nazaninhoseinkhan]

Hi every body,

I want to analyze the results of Cuffdiff.

Cuffdiff has detected 149 significant differential expressed genes.

However, some genes have FPKM=0 (no read has been mapped) or very small FPKM in one condition.

My question is, should I consider those with FPKM=0 or very small FPKM?

Is it biologically possible that one genes does not expressed in one condition and over expressed in another condition?

I am looking forward your comments

Regards

Nazanin