This is a test version of Biostars. For the public version, visit https://www.biostars.org.
CPM function in EDGER

Hi, I have some doubts on cpm function in edgeR package. Given this initial part of the code:

> library(edgeR)
> data_clean<- read.table("unito.txt", header = TRUE)
> group<-factor(c("A", "A", B","B"))
> y <- DGEList(counts=data_clean,group=group)
> y <- calcNormFactors(y)

I obtain different values if I calculate cpm in this way

> cpm <- cpm(y, log = FALSE, normalized.lib.sizes=TRUE)

Or in this one

> cpm <- cpm(y$counts, log = FALSE, normalized.lib.sizes=TRUE)

Why are them different? What does it change? What is the correct way to calculate them? I noticed that if I calculate them with the standard formula "by hand", I obtain the same values as the second formula (y$counts). In the edge manual it seems just to use the y argument.

Thank you, Francesca

edgerr rna-seq cpm
• 11,069 views • link
updated by ATpoint ATpoint • written by francesca3 • francesca3

Please use the formatting bar in edit mode (especially the code option) to present your post better. I've done it for you this time.
code_formatting

Thank you!

• 0 views • link
written by GenoMax GenoMax

Hello francesca3!

It appears that your post has been cross-posted to another site: https://support.bioconductor.org/p/133466/

This is typically not recommended as it runs the risk of annoying people in both communities.

• 0 views • link
written by Ram Ram

1 answer

This is normal and expected. If you use the first chunk the function will use the norm.factors from calcNormFactors to correct the counts for both library size and composition, see this video for the theory of library normalization.

With the second chunk of code you are providing only the count matrix so the function has no norm.factors therefore only correct per-million so for library size. That is why this matches with your by hand calculation.

The first method is the one you should use.

• 0 views • link
written by ATpoint ATpoint

Glad to help!

• 0 views • link
written by ATpoint ATpoint

Log in to answer this question.

More posts like this