BioAnalyzer RIN/rRNA ratio disparity

[JDSwenson]

Hi all,

I am preparing a few dozen libraries for RNA-sequencing. They all come from stingray embryos in early development. To my knowledge, nobody has extracted RNA from any species within the same family, so we're kind of in uncharted waters. Therefore, when I see something odd in my BioAnalyzer traces, it's hard to know what's normal for these animals. I'm particularly vexed by a consistent disparity between the RIN and rRNA ratios.

I am extracting RNA using Trizol followed by a column-based purification and DNase I digestion. After each RNA extraction I am diluting the total RNA anywhere from 1:20 - 1:100 (based on Qubit values) and running it on a BioAnalyzer using an RNA Pico kit (FYI: we could only afford one kit, and we went with the Pico instead of the Nano, due to its ability to also detect low levels of mRNA).

So far I have not had a single sample with a ribosomal ratio > 1.5, but I have also not had a single sample with a RIN < 8.5. In about 50% of the samples, a RIN cannot be calculated because of an 'unexpected ribosomal ratio', which may fall as low as 0.6. When I change the threshold for the ribosomal ratio parameter from 0.7 to 1, thereby forcing a RIN calculation, it is almost always exceptional (see attached traces). Further, all of my samples show similar traces which, besides the low rRNA ratio, show few, if any, other signs of degradation.

I've read conflicting reports about the reliability of ribosomal ratios and RINs, so I bring my inquiry to you knowledgeable folk in the seqanswers community: do you suspect my low rRNA ratios may just be an artifact of dilution or the suggested 70C for 2 min denaturation right before running (I have heard the 28s rRNA can break down a bit during this step)?

More importantly, can I trust my BioAnalyzer traces and proceed with RNA-Seq library prep?

Many thanks for your time,
John

[nucacidhunter]

28s rRNA breakage caused by heating has been reported in insects and if that was the cause then you would expect it in all your samples. Dilution is also less likely cause because the assigned RIN is independent of sample concentration, instrument and analyst but you may have other species in some of your samples with differing rRNA ratios that affects the overall sample rRNA ratios (I do not know how the embryos have been collected and the possibility of contamination). You can find an example of 28s rRNA breakage upon heating and Bioanalyser run in following paper:

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3016993/

I do not see any signs of RNA degradation and I assume you are going to do polyA selection/priming so it is safe to proceed with library prep because:
1- If these samples were run on a gel for QC based on gel image I would have proceed with library prep.
2- The ratios are relatively close in all samples so it does not introduce variability
3- The ratios can even vary in different tissues of the same species

[JDSwenson]

Awesome. Thanks a million, nucacidhunter. I really appreciate the link and the explanation. I am the first on my campus to prepare libraries for RNA-Sequencing, so when I run into a problem, it's helpful to have somebody as kind as yourself who's willing to take the time to help me troubleshoot and explain things I may not have understood.

So, thank-you.

[nucacidhunter]

Your are welcome, happy to help.