qvalue package in R: Can it be used with very small proportion of true alternatives?

Question

[ckhh]

Hi,

I'm trying to use 'qvalue' package in r to work out the true proportion of alternatives.

I get pi0 is >1 with small lambda. pi0.lambda seems to decrease as lambda increases.

In my simulated dataset I have only a very small proportion of true alternatives. (approx 0.001%)

Can this be the reason? Is there any way I can work out the true proportion of alternative test?