How can CAGE and RNA-Seq data complement each other?

Question

[espop23]

Hello,

I am new to these technologies, and would like some more information to distinguish each. In what ways can CAGE and RNA-seq data complement each other? What is unique for each of the technologies?

For instance, given I have some data that supposedly also has matching CAGE data that is mapped and clustered. If these samples match the RNA-Seq samples, how do I know if the CAGE samples can complement the RNA-Seq analysis?

Thanks for any help!

Answer 1

[GouthamAtla]

In CAGE, only mRNA with 5' caps are captured and sequenced.

Usually there will be multiple TSS for a gene and some genes may have multiple promoters. In RNA-Seq you will get to know which genes/transcripts are expressed, because the mRNA is randomly fragmented and sequenced, but that does not tell the which promoter being used. CAGE-Seq on the other hand, sequences the 5' regions of mRNA, which tells you the precise location of TSS at 1bp resolution, which explains you about the promoter usage and promoter architecture.

In short, if you want to study the promoter architecture, you need to use CAGE "like" data. If you are only interested in expression, RNA-seq would be enough. If you have deep enough the CAGE data, you could study both the promoter architecture and the gene expression.

Answer 2

[Chirag Nepal]

There is comprehensive comparison of RNA-seq and CAGE-seq data which was published from FANTOM5 consortium. Check their paper. We had also compared RNA-seq and CAGE in zebrafish, you can check here http://www.ncbi.nlm.nih.gov/pubmed/24002785

In general, both CAGE and RNA-seq are concordant, and each of these technologies have their own advantages, when used in combination is good.

[espop23]

In what way would you typically use the 2 in combination?

[Chirag Nepal]

Specially, when you find alternative intronic promoters from CAGE, you can can ask if the CAGE detected 5'-end is supported by RNA-seq 5'end. You can also correlate how CAGE tags from CDS with RNA-seq. CAGE tags originating in promoters or introns or CDS have different initiator sequence, that can be very useful to make sense of your data. Please read the paper: http://www.ncbi.nlm.nih.gov/pubmed/24002785

Answer 3

[Charles Plessy]

For a visually interactive answer to your question, you can visit an example in our "Zenbu" browser, where CAGE and RNA-seq data from ENCODE are displayed.

Also, there are a wide variety of RNA-seq protocols, so the answer to your question may be very different according to the RNA-seq data you have or plan to produce. CAGE is strand-specific and random primed, which makes it a good tool to detect the promoter of non-polyadenlyated long RNAs, which is desirable when the study aims at covering long non-coding, antisense or enhancer RNAs. Some RNA-seq procedures lack both properties: made from oligo-dT-primed RNA and not distinguishing both strands, they would perform very poorly at this task.