Differential gene expression analysis ?

Question

[sunnykevin97]

Hi

for differential gene expression analysis do we take account only all the common transcripts across the samples or else, we take all the transcripts present in all samples ? please some comments!

Answer 1

[Kevin Blighe]

Hello sunnykevin97,

It can be accepted that some genes / transcripts will have expression values of 0 (zero) in some samples, which may be encoded as NA, depending on your upstream pipeline. There is no justification to remove a gene / transcript that has a value of 0 / NA in just one sample, though.

What you can do is one or both of the following prior to normalisation:

• remove genes / transcripts that have a low mean across all samples
• remove genes / transcripts that have a high rate of being 0 / NA across all samples

If you would like further help, then please be very specific about the programs and methods that you are using so as to avoid confusion and a lengthy thread.

Kevin