PolyA reads are disappeared?

[SOLiDrobino]

Hi at all,
I've done an RNA-seq experiments with SOLiD 4 using a polyA-enriched RNA as the input sample. Interestingly I've found only few reads with a stretch of 0 longer than 20. I was expecting that many reads were made of 0 as a product of the sequencing of the polyA tail of mRNA. Anyone had similar results?
Thank you

[pmiguel]

(Took me at least a minute to get your "stretch of 0 longer than 20" comment. My head just isn't in color space these days, I guess. "0" is the designation for all the homopolymer di-bases: "AA", "TT", "CC" and "GG". So a poly A tail would be a poly 0 tail in color space...)

I think that is a common result. We discussed it at length in this thread:

http://seqanswers.com/forums/showthread.php?t=3195

starting at post 20. (But feel free to read the entire thread. Although it is a bit of a history lesson because Ambion changed the kit to avoid the very issue that hingamp reports.)

If you don't feel like reading the entire thread, you might focus on the post where I lay out a number of possibilities:

http://seqanswers.com/forums/showpost.php?p=11254&postcount=27

My pet hypothesis, in that post, is that the "TTTTT" pentamers end up getting depleted from the pool by sitting on non-ligatable parts of the poly A tail.

I'm sure you could come up with some other, probably more likely, hypotheses.

--
Phillip

[SOLiDrobino]

Dear Phillip,
thank you for your answer. I'm new of this forum and I didn't saw the previous discussion so I started a new discussion. You and other guys said a lot of interesting hypotheses. The idea of a RNase III bias against polyA cutting was my first hypothesis but there are no evidence about this. Nevertheless in my lab we had tried to fragment RNA with different enzymatic reaction so, when I'll get the sequences, I'll report to you if there are some interesting evidences.
The idea of a 3'-blocked end of the polyA for me is not so good; polyA can be very long so if the real 3'-end is blocked, after the fragmentation many other fragments of polyA will had ligation-compatible ends. I have no ideas about your theory of the functional depletion of adaptors with TTTTT protruding end, but my hope is that each adaptor is in large excess respect to the RNA fragments ends.
Now I'd like to know if also with other second generation sequencers there is the same problem (I've never sequenced mRNA with 454 or Solexa), because it could be a problem with the ligation-sequencing chemistry like the dissociation of the probes from a stretch of A (the annealing temperature must be very low)...briefly, I don't know, but I'm very sad because I needed that polyA sequences!!!

[pmiguel]

What enzyme? Unless you are using RNaseIII, RNaseH or RNaseP (which I don't even know a commercial supplier for), then you would end up with 3'-PO4 (or cyclic 2'-3' PO4) and 5' OH, right? So you would have to dephosphorylate/kinase the fragmented RNA ends to make them compatible for ligation.

At that point seems like you might as well use divalent cations and heat.

I guess RNAseH could work, but you would need to really tinker with the protocol because its natural substrate is RNA/DNA heteroduplexes. Also I have been lead to believe that polypurine stretches would be resistant RNAseH digestion. (Hence the functionality of the polypurine tract in retroviral plus strand synthesis.) So that might create a bias for you as well.

--
Phillip

[SOLiDrobino]

We tried to digest our mRNA samples with S1 nuclease and Mung Bean Endonuclease. In both cases we checked the fragmentation with Agilent 2100 Bioanalyzer and with our digestion condition the profiles seems very good. Both enzymes leaves 3'-OH and 5'-PO4 (checked also with RNA oligo ligation, RT-PCR and Sanger sequencing of a digested synthetic RNA).

[pmiguel]

I had no idea either of those enzymes worked on RNA. That is great!

I really hate using RNAseIII -- very strong cleavage bias. The only reason I can think of for Ambion using it are the ends being ready to ligate. Probably much less bias with S1 or MBN.

I think the ligation must be much less efficient than hexamer primed reverse transcription because the starting amounts for that protocol are orders of magnitude higher. 200 ng of ribo-depleted RNA! The Illumina TruSeq RNA allows starting amounts of total RNA as low as 100 ng.

Will be very interested to see your results...

--
Phillip