Hi Christopher

Very relevant subject...

In fact there are 2 main points in your discussion:

1.- RIN: this "golden" number is calculated by a ratio between big and small rRNAs, being an indication of RNA degradation (in an overall sense). miRNAs are extremely stable against nucleases, and RIN cannot be considered at all as a "degradation indication" for those small RNAs. In this sense I agree that RIN is not a good factor to estimate miRNA degradation. You gave us extremely solid arguments to support this fact, specially when you talked about the paraffin embedded samples, where big RNAs are usually a mess and miRNAs are still stable. This is because miRNAs have been considered as a very convenient biomarker for samples collected not in the "best" conditions.

2.- Phenols: that's a different story. I am not too sure that phenol is irrelevant in quantification, since phenol can interfere with the RT reaction performed before miRNA quantification by qPCR.

Hope it helps and happy to continue the discussion.

Thanks for such a good post. Best

Paco 

I agree, as always, with Paco. Basically the RIN value is irrelevant for microRNA analysis in my experience. It becomes an issue when you use non-microRNA normalizers. These non-microRNA normalizers will be degraded and mess up your analysis if used.

If you look at bioanalyzer results one can see that degraded samples have degraded RNA that never ever mixes with the microRNA peak. RNAses, heat etc must have a limit of degradation and will not produce really small fragments that mix with microRNAs. So even purification of this microRNA peak may not contain substantial amounts of degraded ribosomal or other degraded RNA. Does it?

This is pretty amazing. No idea what the biological basis for this is ( as suggested, the enzymes that degrades RNA stops at way above 20 nucleotides. Strange, isn't it!? Check out bioanalyzer results, they are very informative in this regard.

An entirely other issue is contamination with phenol and salts. they will have an effect.

Thank you for your answers! I agree that organic contamination will affect the efficiency of my reverse transcription, but I am wondering if this is also something that can be corrected for for the following reasons:

1. The 260/230 ratios are similar between experimental groups, there is therefore no bias is RT-efficiency, and reduced efficiency will affect each group equally. The mean difference between groups should therefore not change that much, but measured variation (noise) in each group will increase. By increasing group size we can reduced the confidence interval of the mean and therefore increase the power to detect real differences. I therefore argue that the differences we find are true differences, but that the hypotheses we reject may be wrongly rejected.

2. In our material, there was actually no correlation between 260/230 ratio and overall miRNA expression (see figure posted with the question above), which leads me to doubt the extent to which the RT reaction is actually impaired. 

Hi all

Thomas always giving important details !... I agree with the problem with the non-miRNA normalizers, but for this reason you should avoid to use big RNAs to normalize the small ones. You can always use low-abundance tRNAs like lys-tRNA or snoRNAs to normalize if you are not sure what miRNA you can use.

I have no idea also about the reason why the cells have enzymes that degrade RNA to fragments around 20 nt. The size of the catalytic core of these enzymes could be one reason for this fact, but I also agree that there is no clear physiological information about why the cells have such a limitation. There is a clear parallelism with Ago proteins. In this case the size of the catalytic pocket of these proteins is the reason why the miRNA seed sequence has 8 nucleotides and not more. 8 nucleotides is the size of RNA that will fit like a glove in the catalytic pocket of Argonaute proteins. 

Whatever the reason is, in the meantime we have good stable biomarkers... :)

Best

Thanks Paco for the info about the pocket size!  I have to bluntly admit that I never read structural biology papers with Ångström and crystallography. I apparently should.

Christopher, thanks for your insights. My position in regard to "impurities" in RNA solutions is that there is not much that can be done about the degree of degradation of the RNA but one can always re-purify and remove impurities. How much these impurities influence the RT-raction is unclear. Does the solution has to "stink" phenoly to have inhibitory levels of phenol or where is the threshold. Same for salts. Since there is little work for the RT enzyme on microRNAs compared to long mRNAs, one could envision a robustness of microRNA expression analysis that cannot be achieved for mRNAs. Not sure whether there is evidence for it despite the implications this may have for using microRNAS as diagnostic and prognostic indicators. 

Hi Christopher,

With respect to integrity of miRNA, RIN is irrelevant, that's for sure. But I guess in a sample with lower RIN (indicating degraded RNA), one might find a lot of contaminating RNA fragments alongside miRNAs. So, on a whole, stable miRNA population along with degraded RNA fragments can result in a sequencing signature considerably different from high RIN sample.

The ratio of RNA extracted from plasma/serum was raged from 1.8 to 2.01. However, the RIN values were very low and not good at all. Our aim is to sequence the circulating micro RNA in plasma/serum. What is your suggestion if we proceed to sequencing with very low RIN value for such type (plasma/serum) samples?

Thank you,