I'm unfamiliar with false discovery rates and I'm taking a look at the QVALUE program. Using this software, my p -value histogram and Q-plots look pretty well (more or less as in Hedelfank et al 2001), so my p-values seem correctly calculated, but I still don't understand how to choose the FDR threshold. I've seen in one paper (Ribases et al 2009), for example, that they assume a FDR threshold of 15% (instead of the 5% threshold by defect), which corresponds to a significant threshold of P < 0.00191 in this paper.
With my data, the software gives me a Pi0 of 0.7 and none of my q-value is significant, when the fact is that my p-values resist Bonferroni correction (P<0.01), which is a more astringent correction!!!!!!!!!. How can I decide the significant threshold and what am I doing wrong?
I would thank any help
Submit your paper to J Biol Methods today!
Anybody knows the QVALUE program? I need help
No replies to this topic