How many samples ones need to get a high power with RNA-Seq in paired-design experiment?

Question

[jonessara770]

Hello,

I have been asked to design an RNA-Seq experiment for paired samples (cancer/normal) before and after treatment and to suggest the number of samples that needed for the study. I know higher power can be achieved when the number of sample increases but I am not sure what is the threshold beyond a certain replication number, the gain in power is negligible. any suggestion to calculate the number of samples needed?

Thanks Sara

[Santosh Anand]

Some cues http://www.rna-seqblog.com/tag/power-analysis/

Answer 1

[John Ma]

At a minimum 3 biological replicates setting.