Entering edit mode
8.8 years ago
star
▴
350
I have a vector as below , I like to convert q-value to p-value (estimate p-value base on q-value).
data<-c(0.35,0.52,0.12,0.46)
I used below code for converting but I am not sure about it:(I take this code from here)
qvalues<-as.numeric(as.character(data))
convert.qval.pval = function(qvalues) {
pi0 = max(qvalues)
m0 = length(qvalues) * pi0
return(qvalues * rank(qvalues) / m0)
}
qval_pval <- convert.qval.pval(qvalues)
qval_pval
[1] 0.33653846 1.00000000 0.05769231 0.66346154
Also, I like use "q-value" R package and but for this package I have to have a simulated or data-resampled (e.g., bootstrap, permutation) "null statistics", but I do not know how can I make it.
Code for converting using Q_value:
stat <- data
stat0 <- "creat using bootstar" ?!!!
p.testspecific <- empPvals(stat=state, stat0=stat0, pool=FALSE)
What is the question? you are not happy with the conversion or you want to use q-value package with your own data?