[BioC] LIMMA:nestedF

[daniela marconi]

Hi,
I have a question about decideTests with method "nestedF".

As I  saw, this function can't handle with NA values!

How can I realize a nestedF classification with a fit with NA values?

I found two solutions by myself. But I'm not sure about!!!

1) I've tryed to list the function decideTests and then I've created a
new function decide.mine  only

-removing
if (any(is.na(object$F.p.value)))
   stop("nestedF method can't handle NA p-values", call. = FALSE)
from the code

-and rewriting
sel <- p.adjust(object$F.p.value, method = adjust.method) < p.value
sel<-(!is.na(sel))


I've tryed also with:


fit<-lmFit(MA,design,ndups=1)
fit2<-contrasts.fit(fit,cont.matrix)
fit2<-eBayes(fit2)
fit2$F.p.value<-p.adjust(fit2$F.p.value, method = "HY")
results<-classifyTestsF(fit2,p.value=0.001)

which one is correct?
Or otherwise How I can manage, to apply nestedF approach with an
MA-list with NA values??
Thanks in advance
Daniela