unstranded RNA-seq lib for ncRNA prediction

Question

[Sam]

Dear All

I used unstranded RNA-seq for non-coding RNA prediction, in 1st step I removed all sequence which is overlapping (in sense) exons of the reference annotation. so can I consider the remaining reads that map to exonic antisense strand as antisense transcript?

Thanks

[Eric Lim]

Tough. You can infer by going after annotated antisense transcripts and look for exonic regions that aren't overlapped with protein-coding transcripts in the sense direction. Then inspect the quants case by case and design primers to validate the findings experimentally. There's only so much you can accomplish this with unstranded protocols.

Answer 1

[ATpoint]

If you have unstranded library preparation, you cannot make statements about sense and antisense, which is the big plus of stranded preps.

[Friederike]

Figure 6 of Griffith et al., 2015 offers a good illustration of ATpoint's explanation.

[Sam]

but I can determine which strand a read maps to, right? to have a number of reads that map to sense and antisense strand.

[ATpoint]

Yes, you have mapping to both strands but this is because adapter ligation is random and after PCR amplification, the strand-of-origin cannot be identified anymore unlike in stranded preps. Sorry about that but you cannot answer the question you have with this experimental design.

[Sam]

ok, thanks . just in case it's not also possible to distinguish between sense and antisense overlapping transcript via qPCR and primer design, right?

[ATpoint]

There are protocols for stranded qPCR. Still, you might consider preparing a new RNA-seq stranded library. Most sequencing facilities offer all-in service for library prep and sequencing for an affordable price. qPCR can be annoying and in the end only gives you information about a few regions.