To Helicos or Not? Weird bioanalyzer profile: DNA or RNA?

[Dapip33]

I am prepping RNA for Helicos sequencing from bacteria that have been fixed with paraformaldehyde. The RNA is purified with the FFPE kit from Qiagen, quantitated with the nanodrop, and a sample is run on the bioanalyzer. I feel that I surely have some degredation, but what concerns me most is the large peak I have right around the maker, which is 25 bp. I'm worried that this is DNA left over from DNAse I treatment as part of the Qiagen kit. Does anybody here have experience with prepping PFA fixed samples and have any similar experience? I feel like there is no good way to distinguish what this is given that the fragments are so small, making PCR to detect DNA impossible.

[VanessaS]

Eww. Looks very degraded to me.
I don't have any experience prepping PFA samples, nor with prokaryote samples, but if this is to judge the quality of your Total RNA, before library prep I'd not proceed. The RIN is 2.8, which is horrid. It should be around 8-10.
The smaller fragments could be small RNA or just the degraded byproducts.
Maybe someone with more experience with your application will chime in!

[pmiguel]

Hi Dapip33,
Could you convert your baseline from "seconds" to "nucleotides"?
If that peak were degraded DNA, how would that impact you? Most of us will not be familiar with the Helicos sample prep protocols.

--
Phillip

[Dapip33]

To answer Vanessa:

I think that the paraformaldehyde fixation is going to result in a weird, atypical bioanalyzer peak no matter how well the RNA is prepared. Take a look at the PDF file in this link from Duke:

http://bit.ly/ey2pJD

This is for Formaldehyde Fixed and Parafin Embedded tissue samples, which are more extensively treated than my sample, but clearly degradation and weird profiles result from fixation. So we think we can proceed with our RNA, but are worried about DNA contamination, which moves to Phillips question.

Helicos transcriptomics relies on making a single round of cDNA from the RNA sample, poly-A tailing that cDNA, and then sequencing short reads of 25-55 bp. If we have contaminating DNA, even small fragments, those will get poly-A tailed and sequenced along with our RNA and obscure what we want which is the transcriptome.

You can read more about how this is applied at the Dana Farber facility website, which is where we'll do our sequencing:

http://mbcf.dfci.harvard.edu/Helicos/helicos.html

We're fairly confident that we shouldn't have DNA in our sample since we did DNAse treatment, followed by column purification (Qiagen), but without being 100%, we're worried we'll waste our money on the sequencing runs to find out we have DNA there.

[Dapip33]

It would appear that my peak is most likely very small fragments of DNA left over from incomplete DNAse treatment. We'll try a few more things and I'll post back with the results when I get them, but for now, we're assuming what we have is genomic DNA contamination of our admittedly degraded RNA sample. After consulting with some experts at Dana Farber/Helicos we got the following email:


First, we need to be sure these bizarre peaks are reproducible and not some Agilent Bioanalyzer problem. They all seem to be from the same Agilent chip run. I suggest first running these samples in a seperate run. I sometimes see some bizarre peaks too that disappear in other runs. Running an empty lane with the water or elution buffer they use to dissolve their RNAs is necessary to be sure the water or the machine they are d (happens a lot).


If it turns out these peaks are indeed real (to be clear, I am talking about the peak below the 30 second mark on the x-axis), then they are very likely DNA. DNAse1 chops up DNA, but sometimes the digestions are not complete and you get a lot of short DNA fragments. In this case, we ask users to finish RNA isolation following the manufacturer protocol, and then do another DNAse1 digestion (they can use this http://www.ambion.com/catalog/CatNum.php?AM2235) followed by cleaning with a Qiagen RNA clean-up column.

Most scientists do not know isolation kits always have DNA contamination, even if they include DNAse1 digestion. It is always better DNAse1-treat after RNA isolation and running a gel to be sure there are no bizarre profiles. Many publications out there have suffered from this problem and their cDNA sequencing data actually contains genomic DNA contamination

[pmiguel]

Yes, that seems reasonable.

Since the Helicos instrument is true 3rd gen (single molecule sequencing) you have some issues that 2nd gen probably would not. Since 2nd gen requires creation of amplicons with adaptors at both end of the template tricks can be played to select against genomic DNA contamination of the final amplicon pool.

If you could sequence directly from the RNA using an oligo dT primer, the DNA would probably not produce much sequence to contaminate your results.

--
Phillip

[Dapip33]

Yes, that seems reasonable.

Since the Helicos instrument is true 3rd gen (single molecule sequencing) you have some issues that 2nd gen probably would not. Since 2nd gen requires creation of amplicons with adaptors at both end of the template tricks can be played to select against genomic DNA contamination of the final amplicon pool.

If you could sequence directly from the RNA using an oligo dT primer, the DNA would probably not produce much sequence to contaminate your results.

--
Phillip

This is one of the issues we have with bacterial RNA. The mRNA is not poly-adenylated to a great extent, so purifying it away from rRNA is a trick too. Initially this project was intended for microarray analysis, but the Broad Institute's NIAID project that is kind of funding the work discontinued microarray work for RNA-seq, but the Broad itself does not yet have the procedures for bacterial RNA-seq with Illumina technology in place.

Admittedly, I'm not as familiar with the tricks that can be played with 2nd gen sequencers to avoid DNA contamination of libraries, so I should look into that a little more. It's possible that our work will migrate back over to the Broad Institute once RNA-seq is inline over there.