Can Rna Sequencing Measure Absolute Quantities Of Mrna Transcripts?

Question

[Sm123]

Can data from RNA Sequencing (Illumina) be used to quantitatively measure absolute transcript levels? There are some discussion of converting RPKM to transcripts/ cell, of metrics; but still unclear. Has there been any comparision of this data to quantitaive gene expression (such as Sequenom)? Thanks.

Answer 1

[Mikael Huss]

It's not quantitative unless you use a spike-in; a reference with a known concentration. The early RNA-seq paper by Mortazavi et al. (http://www.nature.com/nmeth/journal/v5/n7/full/nmeth.1226.html) is one of few that try to address the question of RPKM vs transcripts per cell. I don't know of any comparison to Sequenom although such may exist.

[Sm123]

thanks for your answer. That '08 paper by Mortazavi et al seemed to suggest that if starting cell numbers and RNA yields are known-- RPKM can be converted into transcripts/ cell. Do you know of others who migt have tested this?

[Gww]

I think there is a huge assumption in their paper that hinges on homogeneity of the cells they isolated their RNA from

Answer 2

[Philippe]

Hi,

i agree in saying that RNA Seq per se does not provide any clue about absolute level of transcript abundance. And, also, absolute levels are still... relative. If you know the initial number of cells you can have an estimation but that also depends on the volume of this cells. An absolute level of n will be finally of a lower importance in cell with a volume v compared with a cell with a volume of v/2.

A good way to have insights about absolute level of expression per cell is also to look at DNA to RNA ratio. I don't know the technical details of such a protocol (I just heard about it). But if you know that the RNA in your sample weight twice more than the amount of genomic DNA in the same sample then you can normalize your levels of RNA per bp (or Mb) or DNA.

Still you have to be cautious because when you extract your RNA you don't extract all the content of the cells, some RNA (but also DNA) is lost and it is hard to quantify how much exactly. So be aware that even such approaches (DNA/RNA ratios, counting cell numbers) will still be a proxy of the absolute level of expression, not an ultimate value.

Answer 3

[Jeremy Leipzig]

RNA-Seq has a high correlation with qRT-PCR (R2>.9), which I believe is the second part of Sequenom's MASSarray technology, the first being MALDI-TOF Mass Spectrometry.