I would like to ask the difference between strand-specific and not strand-specific dataset.

As far as I know, strand-specific data means that we know which strand the transcript is from.

I do not have biological background. Please confirm whether it is correct. If we have a transcript, which is from sense strand, when RNA-seq is producing reads, is that first the cDNA is synthesised. Then this cDNA is used for PCR to amplify the sample? Then the reads generated could be from both strands of the original DNA?

For strand-specific protocols, what is different?


Follow up.

Please correct me if I am wrong. There are multiple protocols to produce strand-specific RNA-seq libraries. The basic process is like:

  1. Get the RNA;
  2. Get its cDNA;
  3. Somehow mark the cDNA as sense or antisense when amplifying (PCR?) (here comes the differences between different protocols);
  4. Then REMOVE all antisense (or sense) cDNAs;
  5. Read the reads from clean cDNA library.

The result is that, the reads from this RNA can be used to assemble the sense cDNA. And for not strand-specific libraries, using the reads would be able to assemble both the anti-sense and sense cDNA.

Am i right about this problem? Thanks.


1 Answer 1


It may be worth checking out the paper "Comprehensive comparative analysis of strand-specific RNA sequencing methods".

The most common techniques sequentially ligate different RNA adapters to the 5' and 3' ends of each RNA molecule prior to cDNA synthesis. You will end up with an RNA molecule that looks like this (where A and B are the two adapters):


These adapters are then used for cDNA synthesis and library preparation. Finally, since each adapter is ligated to a specific end of the RNA the library can be generated in a stand-specific fashion; ie. the reads derived from the antisense strand should align to the antisense strand of the genome.

** Note that there are other variations of this technique but Illumina's protocol is one of the most popular.

  • $\begingroup$ Thanks, @GWW. Does it mean that Illumina is going to provide two libraries (one antisense and one sense) to users? Or they just mark every read which sense it is from? When giving reads to customers, have they removed the adapters beforehand? Or users need to remove by themselves? $\endgroup$
    – Joy
    Commented Apr 25, 2012 at 15:54
  • $\begingroup$ By the way, I checked the paper... But cannot find much regarding the post-processing of the data... $\endgroup$
    – Joy
    Commented Apr 25, 2012 at 15:56
  • $\begingroup$ You mean aligning the reads? $\endgroup$
    – GWW
    Commented Apr 25, 2012 at 16:09
  • $\begingroup$ I mean, you said "the reads derived from the antisense strand should align to the antisense strand of the genome", so what would Illumina provide? If they just provide whatever reads, that is not strand-specific, right? They must filter out sense or antisense reads, or they may mark each read as whether it is from sense or antisense strand. Am I right? $\endgroup$
    – Joy
    Commented Apr 25, 2012 at 16:12
  • 1
    $\begingroup$ Yes, assuming that the libraries were prepared using a directional library. $\endgroup$
    – GWW
    Commented Apr 25, 2012 at 17:47

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .