I am looking inside the DESeq2 code to learn how it works.

I can't find any definition for the function "normalizationFactors".

The reason is that when I run :

normalizationFactors.DESeqDataSet <- function(object) {
  # Temporary hack for backward compatibility with "old" DESeqDataSet
  # objects. Remove once all serialized DESeqDataSet objects around have
  # been updated.
  if (!.hasSlot(object, "rowRanges"))
    object <- updateObject(object)
  if (!"normalizationFactors" %in% assayNames(object)) return(NULL)
  assays(object)[["normalizationFactors"]]
}

setMethod("normalizationFactors", signature(object="DESeqDataSet"),
          normalizationFactors.DESeqDataSet)

setReplaceMethod("normalizationFactors", signature(object="DESeqDataSet", value="matrix"),
                 function(object, value) {
                   stopifnot(all(!is.na(value)))
                   stopifnot(all(is.finite(value)))
                   stopifnot(all(value > 0))

                   # Temporary hack for backward compatibility with "old"
                   # DESeqDataSet objects. Remove once all serialized
                   # DESeqDataSet objects around have been updated.
                   if (!.hasSlot(object, "rowRanges"))
                     object <- updateObject(object)
                   # enforce same dimnames
                   dimnames(value) <- dimnames(object)
                   assays(object)[["normalizationFactors"]] <- value
                   validObject( object )
                   object
                 })

I get the following error:

Error in setMethod("normalizationFactors", signature(object = "DESeqDataSet"), : no existing definition for function ‘normalizationFactors’

Should I define a function like:

defining a function like is OK?

normalizationFactors <- function(object){
             normFactors <- matrix(runif(nrow(object)*ncol(object),0.5,1.5),
                                             ncol=ncol(object),nrow=nrow(object),
                                             dimnames=list(1:nrow(object),1:ncol(object)))

             normFactors <- normFactors / exp(rowMeans(log(normFactors)))
             normalizationFactors(object) <- normFactors
}

normalizationFactors is a generic defined in AllGenerics.R

Defining methods and generics for S4 classes is covered here:

http://adv-r.had.co.nz/S4.html

I’m not sure what advice to give because I don’t know if you want to write your own custom function just for personal use (in which case you should use a name that doesn’t conflict with existing methods) or if you are learning how R/Bioconductor packages are written.