q value doubt

[a-espuela]

Hi,

I have been using the qvalue R package and I came across an unusual behaviour. My input is a list of p-values and what I want to do is to adjust for multiple comparisons.

I have used the qvalue function, to extract afterwards the q-values from the generated object (qvalues and lfdr).

For most of my analyses the q-values are bigger than the original ones (which I think is the normal output) but for one particular analysis, Storey's q-values and lfdr are smaller.

I am wondering if this is normal. Besides, are there any assumptions that this type of approach has that I should be concern of? I have checked the distribution of p-values and it seems OK.

Thanks in advance, Antonio

[jdstorey]

https://github.com/StoreyLab/qvalue#frequently-asked-questions