Total counts(TC) normalization by R

Question

[KVC_bioinfo]

Hello All.,

I am trying to find an R package or R script which can perform total count (TC) normalization in R?

Could someone help me out here?

[Devon Ryan]

The normalizations provided by edgeR and DESeq2 are more robust, why not use them instead?

[KVC_bioinfo]

I want to perform it because i want to compare different normalization methods.

[Kevin Blighe]

Already been done, I'm afraid, mi amigo: A comprehensive evaluation of normalization methods for Illumina high-throughput RNA sequencing data analysis.

Total count normalisation refers to a few different types, such as FPKM and RPKM, which have been shown to be inadequate. They are mentioned in that manuscript.

As per Devon, EdgeR and DESeq's normalisation strategies are more robust; however, ironically, it appears that a fpkm() function comes with DESeq2: https://www.rdocumentation.org/packages/DESeq2/versions/1.12.3/topics/fpkm

[ATpoint]

In case you want TPM (either using the sum of the column as read count or the total count of the BAM file) use this:

# Normalize for gene/region length:
x <- counts.mat / gene.length

# Normalize to read depth (here depth = colSums of the count matrix):
tpm.mat <- t( t(x) * 1e6 / colSums(x) )

[Devon Ryan]

EdgeR has one too, though these are more robust since they use the normalized counts (i.e., they're not what one would get with a naive algorithm).