Dear all,

I'm trying to use LVSmiRNA to normalize a total of ~1700 (agilent) samples. The problem is that the step when estVC is called takes lot of time.

The vignette of the package propose to use multicore package but this packages is deprecated and CRAN suggest to use parallel.

Hence I just changed one line from the proposed code from the LVSmiRNA's vignette to run estVC. This code follows:

require(parallel)
options(cores=16)

require(LVSmiRNA)
require(limma)
MIR <- read.mir(mirna_files, path=mirna_path, verbose=TRUE)

colnames(MIR$E) <- colnames(MIR$Eb) <- samples_names
class(MIR)<-"EList"

MIR.RA <- estVC(MIR, verbose=TRUE)

I just expected to get a different output from estVC when parallel is loaded and cores option is set to a large number than the one I obtained:

using lapply

May I should indicate something else to estVC in order to run it faster than sequentially? Otherwise, may you know a faster way to apply this normalization to a large set of samples?

The last version of the package allows to include a "subsample" to the estVC function. For example:

MIR.RA <- estVC(MIR, subsample=sample(1:50))

With this, computational time is reduced.