RIN issues & Nanodrop/Bioanalyzer disparities

[marymac]

Hi All,
I am at my wits end and was hoping someone might provide some insight. I am working on a transcriptomic project with microfilariae (immature filarial nematodes) and have been using Trizol and the Zymo Research Clean & Concentrate kit. I have been struggling on a few fronts:

1) When I first look at my RNA on the Nanodrop, everything looks nearly perfect, 260/280 consistently higher than 1.85, 260/230 pretty good as well, with high concentrations (500+ ng/ul), peaks obviously look good

2) When I send my samples off for the Bioanalyzer, I am consistently getting large differences in concentration both higher and lower, but mostly lower
Example: Nanodrop=457 Bioanalyzer= 66 or Nanodrop= 619.6 Bioanalyzer=77

3) Bad RIN scores: I am fully aware that I likely have degraded samples, but I was wondering if the Bioanalyzer might have an issue with the sample type itself. These parasites went through a mosquito and into a gerbil before I collected them, so I am wondering if it might be like the issues we see with insect or fungi samples? I have been getting lots of 2-3 scores and only a handful higher than 5 (after 50+ sample submissions with 4 different treatment groups).

I would appreciate any feedback you guys might have, and if this is not the correct forum or you suggest cross-posting, just let me know

Thanks,
Mary

[kmcarr]

Nanodrop quantitation is notoriously unreliable for anything other than highly purified nucleic acids. To accurately measure the RNA concentrations in your samples you need to use a dye-binding fluorometric assay (e.g. Qubit, picogreen).

[adam.geber]

Hi Mary,

1 & 2 are not inconsistent with each other -- the Nanodrop is known to heavily overestimate concentration, especially in the presence of contaminants. Have you checked your samples for residual gDNA contamination? Clean Trizol preps should get rid of most DNA but you can attempt to amplify part of the microfilarial genome to confirm the absence of contamination. Alternatively, if you have access to a fluorometer like the Qubit you can use nucleic acid-specific assays to get a more accurate estimate of your RNA concentration. If your pipetting is good and the rest of the Bioanalyzer chip runs fine I would trust its measurements of concentration above the Nanodrop.

Regarding 3, how are your samples collected/stored before and after extraction? Do you expect there to be mammalian host contamination on the cellular level? Have you checked your RNA integrity before and after freezing? RIN algorithms are optimized for canonical organisms so there may be discrepancy in the scoring, as you mentioned. Also, as I described here, Trizol extractions can lead to a dampening of the 28S band signal and a lower RIN score overall. Can you try extracting your samples with another method to compare the integrity?

[marymac]

I washed all of my samples 3 times to remove gerbil material and stored them in RNAlater. I did the same project with adult worms and the RNeasy kit worked well, but it seems to be pitiful with Mf, leading me to use Trizol and clean & concentrate.

I'm starting to think that the RNAlater might be part of the issue- I extracted RNA from similar samples that were stored (frozen -80 ) in Trizol and they consistently gave better RINs, but alas they were not part of my experiment set. Mf have very tough sheaths and cuticles, so I'm wondering if the RNAlater didn't saturate them quick enough to prevent degradation.

Interestingly enough, I haven not been able to easily find a paper that did RNAseq with Mf in which they report RIN scores at all (as in we used RIN scores of 8+, etc, which is the goal for my study). Most papers just ran gels and went with that. This is what led me to wonder if there was something particularly tricky about my sample type.

As it stands, my PI and I have decided that samples with RINs of 6+ are what we are going to go with, I'm just running out of samples to extract from.

[adam.geber]

This certainly wouldn't be the first time that I've seen RNAlater complicate extractions downstream -- it's certainly possible that it's not penetrating the cuticles. One of the best options with regard to RNA integrity is to homogenize fresh samples in Trizol before storing.

Depending on your RNA enrichment procedure(s) downstream the "low" integrity may not be an issue at all.

[marymac]

This certainly wouldn't be the first time that I've seen RNAlater complicate extractions downstream -- it's certainly possible that it's not penetrating the cuticles. One of the best options with regard to RNA integrity is to homogenize fresh samples in Trizol before storing.

Depending on your RNA enrichment procedure(s) downstream the "low" integrity may not be an issue at all.

Could you elaborate a little? We are planning on using the NextSeq Platform with the TruSeq kit like we did with adults. We have a bioinformatician working with us from there.

I've been kicking myself all week for the 20/20 I have now about how to store the Mf, but the experiments are done, Mf are in what they are in, and I've gotta work with what I've got

[cmbetts]

Could you elaborate a little? We are planning on using the NextSeq Platform with the TruSeq kit like we did with adults. We have a bioinformatician working with us from there.

I've been kicking myself all week for the 20/20 I have now about how to store the Mf, but the experiments are done, Mf are in what they are in, and I've gotta work with what I've got

I think he means that if you were rRNA depleting, rather than polyA selecting, the RIN score is far less important because it doesn't introduce 3' bias. However, that's unlikely to be an option for you since the rRNA depletions are sequence specific and generally only commercially available for very common organisms (human, mouse, rat).

[adam.geber]

I think he means that if you were rRNA depleting, rather than polyA selecting, the RIN score is far less important because it doesn't introduce 3' bias. However, that's unlikely to be an option for you since the rRNA depletions are sequence specific and generally only commercially available for very common organisms (human, mouse, rat).

Exactly. One of my collaborators uses Terminator exonuclease for microfilarial rRNA and we see a reduction in ribosomal sequences that is essentially proportional to sample integrity. polyA enrichment might be your best option unless you want to go through the validation of an RNase H-based rRNA probe set.

Out of curiosity, did you run into any of these issues with your adult worm samples?

[marymac]

I had some issues getting good RNA out of male samples at first, but that was because they are smaller and I needed more of them to get good yields. It was definitely a learning curve. What bothers me is that the RNeasy Kit that worked beautifully for adults is absolute rubbish for Mf, whereas Trizol seems to work much better for Mf than it did for adults. I had assumed that RNeasy worked better than Trizol for adults because of a higher chaotropic salt concentration, and would carry over for Mf, but that is not what I've seen.

After people suggested a non-Trizol method I went back and tested Mf in the RNeasy kit again and the results were pretty bad in terms of yield (<40 ng/ul on Nanodrop), but I guess I'll send them off to the Bioanalyzer to see if they give me better RINs. Maybe I'm just being greedy with wanting both a high yield and a high RIN, but if I have to sacrifice one for the other, I'll do it and cross my fingers that I have enough starting material for library prep.

Another thing I noticed scanning through my numerous Bioanalyzer reports is that many of the samples with really bad RINs (~2) had RNA concentrations that actually agree very well with Nanodrop.

[adam.geber]

marymac, one obvious question: which version of the RNeasy kit were you using? I find that the RNeasy Micro kit worked far better at retaining small amounts (<100ng) of RNA than the Mini version.

I can't think of a reason why yield and integrity would be radically different unless you're doing some kind of homogenization that's shearing your RNA. How are you breaking down the Mf sheaths/cuticles?

Also, make sure you have access to a Qubit or comparable fluorometer before starting the TruSeq prep! The Nanodrop is consistently overestimating your sample concentration and you'll want to make sure your input for polyA selection is in the range 0.1-1ug. Likewise for checking your cDNA library yields after adapter ligation and PCR enrichment!

[marymac]

I used the regular RNeasy Plant Kits (regular reagents RLT +Bme, RW1, RPE) for adults and a few Mf samples. I found that to get decent yields for Mf (in Nanodrop's eyes), samples are put in Trizol (about 3 ml per sample), flash frozen and crushed with a mortar and pestle (while still frozen). Then to ensure homogenization, I pass the samples through 25 g and 30 g needles several times, as a former post-doc showed me. I then add chloroform, spin and proceed with the Clean & Concentrate kit.

In terms of library prep- I have always used the Bioanalyzer's concentration as the determining factor for the amount of RNA going into the TruSeq prep kit, even before I realized how off the Nanodrop can be.