Hi,

I'm trying to use derfinder for my RNA-seq data and am having trouble decoding the normalisation steps. In the userguide example the data is already normalised to total number of mapped reads using:

system.time(fullCov <- fullCoverage(files = files, chrs = 'chr21',
    totalMapped = rep(1, length(files)), targetSize = 1))

I can successfully upload data following the guide and have named it "files". I have also used getTotalMapped() for all samples and am using these numbers in totalMapped, but what do I put in targetSize? What does this parameter do? In this example I have put the mean of all the library sizes.

reads=c()
for(i in 1:24){
  reads[i] <- c(getTotalMapped(files[i], chrs = NULL))
  print(reads[i])
}

print(i)

system.time(fullCov <- fullCoverage(files = files, chrs = c('chr1','chr2','chr3','chr4','chr5','chr6','chr7','chr8','chr9','chr10','chr11','chr12','chr13','chr14','chr15','chr16','chr17','chr18','chr19','chr20','chr21','chr22','chrX','chrY','chrM'),
                                    totalMapped = reads, targetSize = mean(reads)))

Cheers,

Helene

Hi Helene,

targetSize is the target library size. For example, 80 million reads which you could pass as 80e6. Set it to a number you can interpret later. For example, at LIBD for one of our projects we aimed to sequence 80 million reads. You could set it to 1 million reads to get counts per million (CPM). Keep in mind the targetSize later on when you choose the cutoff for the expressed regions approach.

Best, Leonardo