Low bacterial RIN values

[Tom_C]

Our sequencing center has informed us that our RNA quality does not pass their guidelines, and we are not certain how to proceed. The center recommends using RNA with a RIN value of 8 or greater, however I have read that some people suggest never using samples with a RIN below 9.7, while others state that different species are fine down to RIN values of 3 or 4. All eight of our samples fall in a range of values from 5.5 to 6.3, and the profiles are quite similar. I have attached a representative Bioanalyzer profile from a sample with an RIN of 6.3.

Can anyone give me further insight on the quality of these RNA? Does this look like typical bacterial RNA, (Enterobacteriaceae) or is it too far gone? We had gone to great lengths to ensure there were no DNA contamination in the samples, involving multiple rounds of extended DNase treatment.

Additional info:

RNA purification with Qiagen RNeasy Protect Bacteria mini kit
DNase treatment with DNase I amplification grade, cleaned up with Qiagen MinElute RNA kit
Second DNase treatment followed by cleanup and elution into water

RNA destined for Illumina HiSeq 2000 100bp PE sequencing

[pmiguel]

RINs in the 5.5-6.3 range would mainly be problematic for poly A+ libraries. Which, I presume, you would not be using. You might see slightly higher rRNA amounts, depending on your depletion method, with slightly degraded RNA, but otherwise I don't see a problem.

--
Phillip

[Tom_C]

Thank you for the response Phillip.

You are correct about the poly-A, we plan to first rRNA deplete using Ribo-Zero (for gram-negatives) before constructing a stranded TruSeq library.

We have chosen to not go ahead with these samples and instead try to extract new RNA of higher quality. I was wondering what the high background in the above BioAnalyzer trace is comprised of. Does this 'hump' contain degraded rRNA molecules, as evident from the decreased 23s peak?

[pmiguel]

Thank you for the response Phillip.

You are correct about the poly-A, we plan to first rRNA deplete using Ribo-Zero (for gram-negatives) before constructing a stranded TruSeq library.

We have chosen to not go ahead with these samples and instead try to extract new RNA of higher quality. I was wondering what the high background in the above BioAnalyzer trace is comprised of. Does this 'hump' contain degraded rRNA molecules, as evident from the decreased 23s peak?

Yes, presumably.
My impression is that ribo-zero has heavy probe coverage of the rRNA targeted. So, you could probably have gotten away with using those slightly degraded RNAs. But if you can get less degraded RNA, that would be better.
--
Phillip

[kerplunk412]

Thank you for the response Phillip.

You are correct about the poly-A, we plan to first rRNA deplete using Ribo-Zero (for gram-negatives) before constructing a stranded TruSeq library.

We have chosen to not go ahead with these samples and instead try to extract new RNA of higher quality. I was wondering what the high background in the above BioAnalyzer trace is comprised of. Does this 'hump' contain degraded rRNA molecules, as evident from the decreased 23s peak?

I agree that the hump is likely degraded rRNA. This is often accompanied by a "peak" of degraded products around 100 nt, but I suspect you aren't seeing that here due to your extraction method. This also indicates that the RNA degradation is happening before extraction.

rRNA depletion supposedly works well even from pretty degraded RNA samples (see this very good paper), so if you are unable to obtain higher quality RNA you will probably be fine going ahead with those samples.

[Tom_C]

So we have re-extracted RNA from our samples and the Bioanalyzer results look good, all samples fall between RINs of 8.1 - 9.1 (representative Bioanalyzer profile of RIN 8.8 attached).

However after DNase I treatment we are still detecting genomic DNA contamination, which shows up at around 20 cycles during qPCRs. Our ability to extract high quality RNA seems to be inverse to our ability to fully remove contaminating DNA.

Any input on how much gDNA contamination is too much when samples are destined for differential RNA-seq?

[kerplunk412]

gDNA contamination isn't much of a problem for RNA-Seq, as the contaminating gDNA is too large to be well represented in the final library (or to cluster efficiently on the flow cell if it does make it into the library), so I don't think what you are seeing will be a problem.

[huguesparri]

gDNA contamination isn't much of a problem for RNA-Seq, as the contaminating gDNA is too large to be well represented in the final library (or to cluster efficiently on the flow cell if it does make it into the library), so I don't think what you are seeing will be a problem.

We already faced issues using depleted RNA contaminated with DNA. As we didn't perform the depletion ourselves (we only built the libraries), I don't know what amount of DNA was present in the sample but it definitely had an impact on the sequence we produced.

[nucacidhunter]

So we have re-extracted RNA from our samples and the Bioanalyzer results look good, all samples fall between RINs of 8.1 - 9.1 (representative Bioanalyzer profile of RIN 8.8 attached).

However after DNase I treatment we are still detecting genomic DNA contamination, which shows up at around 20 cycles during qPCRs. Our ability to extract high quality RNA seems to be inverse to our ability to fully remove contaminating DNA.

Any input on how much gDNA contamination is too much when samples are destined for differential RNA-seq?

Sample BA profile does not indicate presence of gDNA. You might need to review the qPCR setting and results.

[pmiguel]

Just do another DNAse treatment followed by a clean-up, if you are concerned.

How much of a problem the DNA will be depends on your procedure for generating cDNA. As kerplun412 points out, the genomic DNA itself will probably be too long to create amplicons out of directly. However there is usually an RNA denaturation/fragmentation step prior to first strand synthesis and this will denature some of the DNA, making it accessible to first strand cDNA primers (usually random hexamers).
Reverse transcriptase will probably extend these primers unless it is in the presence of Actinomycin D. But, even then, you would have a shorter stretch of nascent strand DNA annealed to the otherwise single stranded, and long, genomic DNA. Again, not a substrate for ligation to adapters yet.

But during 2nd strand synthesis primers might anneal. Less likely here as few protocols have heat denaturation steps post 1st strand synthesis. Instead they tend to rely on the RnaseH activity of the reverse transcriptase to degrade/displace the RNA template strand prior to/during 2nd strand synthesis. RnaseH should do nothing to the genomic DNA.

So, all in all, I would concur with kerplunk412 -- probably not an issue for you.

But, again, do another DNAse treatment/clean-up, if this is a concern.

--
Phillip