Rpkm Calculation For Genes

Question

[siddharth.sethi5]

Hi everyone

I understand the RPKM fomula is as follows:

C = Number of reads mapped to a gene

N = Total mapped reads in the experiment

L = exon length in base-pairs for a gene

Equation = RPKM = (10^9 * C)/(N * L)

I have the counts ( from HTSeq) and transcript length (retrieved from Ensembl API) for each gene.

My question is, for the total mapped reads(N), should I be counting the reads only in the exons for all genes? If thats right, can I just add all the gene counts from HTSeq output to get the total mapped reads.

OR
the total mapped reads will be ALL the mapped reads in the BAM file?

[O.rka]

Can somebody confirm that this is the same equation as http://www.rna-seqblog.com/rpkm-fpkm-and-tpm-clearly-explained/? Why is 10^9 different that 1e6? CORRECTION: nevermind, I didn't see the 1e3 incorporated in one of the other eqns

Answer 1

[JC]

Well, maybe you don't need RPKM any more: "Measurement of mRNA abundance using RNA-seq data: RPKM measure is inconsistent among samples". Wagner GP, Kin K, Lynch VJ. _Theory Biosci._ 2012. PubMed

[Damian Kao]

That's an interesting paper. Here is a link to the full text: https://springerlink3.metapress.com/content/18284k158v887k57/fulltext.html

[Istvan Albert]

Nice blog post on this Damian: RPKM measure is inconsistent among samples. Of course this begs the question, just how many of the papers using RPKM are not quite correct?

Answer 2

[Ali]

First of all, it does not matter if you are going to perform differential gene expression analysis or something like that.

Note that N is common in all genes, so either of the options will only scale your numbers.

If you are focusing only on the expression level of the genes, perhaps a good way is to set N as the number of reads mapped to the exons, to have the sum of C for all genes equal to N.

If you are interested in miRNAs, ncRNAs and so on, you may keep N as all of the mapped reads.