PDA

View Full Version : Poly-A or Ribosomal RNA depletion


NicoBxl
12-06-2011, 03:25 AM
Hi,

For RNA-seq experiment. Is it better to use poly-A or ribosomal RNA depletion library kit ? Is the sensitivity better in poly-A (for mRNA detection) ?

Thanks,

N.

clissold
12-06-2011, 04:11 AM
Hi there,

I've tried many different kits to optimise this protocol and the most effective one to date is the Epicentre Ribozero kit.

Hope this helps.

NicoBxl
12-06-2011, 04:33 AM
Do you have good correlation results with poly-a librairies ? And do you detect low expressed transcripts ?

madseq
12-06-2011, 06:06 AM
Hi NicoBxl, Iīm interested in your question too...

Everytime I use Ribozero, which is very effective at rRNA removal, I wonder how many mRNAs are being lost and if we are creating a lot of bias there. When we started using Ribozero we almost threw the first sample away because we thought that after rRNA removal we didnīt have RNAs anymore...

Anyone has done correlation studies?

pmiguel
12-06-2011, 06:10 AM
If you have intact total RNA and no special interest in non-polyadenylated messages, then poly-A RNA purification is probably going to give you results more in line with what you expect. That is, expression profiles of the exons of the (mostly protein coding) genes of your organism.

If your RNA is not so intact, and/or you are interested in non-polyadenylated messages and/or just concerned that factors such as polyA tail length might bias your RNAseq data set -- then I would also recommend ribo-zero. But also, unless you know yourself to be a bench ninja, replace any step in the protocol calling for ethanol precipitation with a good column based approach instead. (E.g. zymo columns).

Let me just amplify that last point: ethanol precipitations of less than 1 ug of RNA or DNA should only be attempted by bench scientists with >10,000 hours of experience with non-kit-based molecular biology or those with access to extensive psychological (grief) counseling resources.

--
Phillip

cascoamarillo
12-07-2011, 09:49 AM
This reminds me a paper I've recently read:
http://www.ncbi.nlm.nih.gov/pubmed/21737426
They compare poly-A selection, Ribominus and another RNA-SEQ kit (from Nugen).

btw; Phillip, I think you exaggerate a little with the bench hours and the ethanol precipitation :)

protist
12-07-2011, 10:10 AM
There was a recent paper http://www.plosone.org/article/info:doi/10.1371/journal.pone.0027288 where they compared RiboZero & RiboMinus depeleted material to each other and to published polyA selected RNAseq data - take home message from their point of view was "We show for the first time, that ribo-depleted RNA-Seq produces reliable coding and non-coding gene expression data and is highly reproducible between different sequencing locations and when using different ribosomal RNA depletion strategies"

With the addition of a carrier - eg glycogen or linear polyacrylamide nucleic acid precipitation of amounts less than 1 ug is not that problematic. The main issue for students or the uninitiated is that the pellet especially when glycogen is used is very loose and care must be taken or it is easily sucked-up when removing the ethanol.

ETHANol
12-07-2011, 10:16 AM
I was about to order a Ribo-zero kit since I will be working with some somewhat degraded RNA and saw Epicentre has a Ribo-Zero rRNA Removal Kit (human/mouse/rat) and a new Ribo-Zero Gold Kit (human/mouse/rat).

Is the new kit better like the marketing says?

BTW, I consider myself to be a full bench Ninja with permanently distorted pipeting thumbs from too many hours in the lab and resuspended endless invisible pellets, and I can say ethanol precipitation of less then 1 ug of nucleic acid without carrier or contaminant is not reliable. Use a spin column, that's what anybody that has taken a look at technology post 1990 would say.

protist
12-07-2011, 10:31 AM
Can't comment on the Ribo-Zero gold kit but I would say that so far for us the RiboZero kits have lived up to the marketing hype. I have used their Gram Positive and Gram negative kits for Beasties whose rRNA depletion in the libraries was previously poor - as tested with MicrobeExpress and RiboMinus. For one of the organisms the rRNA depletion has gone from <30% to >80% when switching from RiboMinus to RiboZero Gram positive. The kits are more expensive but the extra reads and greater multiplexing capabilities we get make it worth it.

pmiguel
12-07-2011, 11:41 AM
With the addition of a carrier - eg glycogen or linear polyacrylamide nucleic acid precipitation of amounts less than 1 ug is not that problematic. The main issue for students or the uninitiated is that the pellet especially when glycogen is used is very loose and care must be taken or it is easily sucked-up when removing the ethanol.

I think what you are trying to say is "you must become one with the pellet".

(BTW, I have seen at least one protocol that warns that glycogen was inhibitory to some down stream reaction. Also, I would worry that the glycogen was purified from some biological source and might contain RNA or DNA from that source...)

--
Phillip

pmiguel
12-07-2011, 11:46 AM
I was about to order a Ribo-zero kit since I will be working with some somewhat degraded RNA and saw Epicentre has a Ribo-Zero rRNA Removal Kit (human/mouse/rat) and a new Ribo-Zero Gold Kit (human/mouse/rat).

Is the new kit better like the marketing says?

BTW, I consider myself to be a full bench Ninja with permanently distorted pipeting thumbs from too many hours in the lab and resuspended endless invisible pellets, and I can say ethanol precipitation of less then 1 ug of nucleic acid without carrier or contaminant is not reliable. Use a spin column, that's what anybody that has taken a look at technology post 1990 would say.

I thought the difference was that they added oligos to pull out mitochondrial rRNA as well.

--
Phillip

rflrob
12-07-2011, 03:01 PM
I think what you are trying to say is "you must become one with the pellet".

(BTW, I have seen at least one protocol that warns that glycogen was inhibitory to some down stream reaction. Also, I would worry that the glycogen was purified from some biological source and might contain RNA or DNA from that source...)

--
Phillip

We use glycogen carrier in our lab for TRIzol purifying 10-100ng of RNA with some regularity, and this has worked well for Illumina RNA seq reagents from the last couple years (i.e. both the old RNA-seq and new TruSeq kits). At those levels, you are basically only seeing the pellet, but it's not too hard to use a Bioanalyzer with the Pico chips to see how well you've done.

protist
12-07-2011, 03:49 PM
Phillip is correct they have added the extra mt rRNA removal to the gold kit.

With regard to the glycogen question we have always used Ambion (guaranteed DNA free) and hundreds of Euk/bac libraries later we have never seen contamination come through. However we did have a collaborator a couple of years back who used salmon sperm in their ChIP protocol & their first ChIPseq results were not pretty! :eek:

As in the main we are still doing 'home brew' strand specific libraries we routinely precipitate our fragmented RNA in the presence of glycogen and have had few problems. Recently for the bacterial libraries we have switched to Zymo RNA (5 ug) Clean&Concentrate columns at the fragmented RNA concentrating stage. We have also used them post the DNase-ing (25 ug variety) for some bacterial RNAs.

NicoBxl
12-09-2011, 12:40 AM
With Ribo zero, is there a size selection step ?

protist
12-09-2011, 12:58 AM
RiboZero is used to rRNA deplete total RNA - there is no size selection involved. Briefly, total RNA is mixed with microspheres & a rRNA removal solution which binds the rRNA and rRNA depleted material is eluted using centrifugation and columns provided in the kit.

NicoBxl
12-09-2011, 12:59 AM
RiboZero is used to rRNA deplete total RNA - there is no size selection involved. Briefly, total RNA is mixed with microspheres & a rRNA removal solution which binds the rRNA and rRNA depleted material is eluted using centrifugation and columns provided in the kit.


So In the sequencing results, there will be also miRNA sequences (and maybe pre-miR sequences ? )

protist
12-09-2011, 01:45 AM
So In the sequencing results, there will be also miRNA sequences (and maybe pre-miR sequences ? )

There could be but it would be dependent on how you isolated your total RNA - if you used an isolation method that preserved all the RNAs. Ther is a nice paper discussing rRNA depletion strategies and whole transcriptome analysis:http://www.plosone.org/article/info:doi/10.1371/journal.pone.0027288. If yourRNA deplete rather than polyA select non coding RNAs will be more likely to be preserved in the subsequent library. However if you are specifically looking for miRNAs an alternate small RNA specific protocol would be the preferred option.

ETHANol
12-09-2011, 01:48 AM
As in the main we are still doing 'home brew' strand specific libraries we routinely precipitate our fragmented RNA in the presence of glycogen and have had few problems. Recently for the bacterial libraries we have switched to Zymo RNA (5 ug) Clean&Concentrate columns at the fragmented RNA concentrating stage. We have also used them post the DNase-ing (25 ug variety) for some bacterial RNAs.

Would you mind sharing your protocol as Illumina is completely backordered on their TruSeq RNA kits.

NicoBxl
12-12-2011, 05:32 AM
with Ribo-Zero, is the poly-A tail detectable ?

protist
12-12-2011, 05:53 AM
For most RNAseq protocols you fragment the RNA and use random hexamers to prime the 1st strand cDNA. The polyA component of your total RNA will still be there after rRNA depletion, the depletion kits only remove small and large subunit rRNA and 5S or 6S RNA and in the case of some kits (RiboZero) the mitochondrial rRNA. What should be left after the depletion is a mix of mRNA (polyA in eukaryotes) and small RNAs.

NicoBxl
12-12-2011, 11:50 PM
Ok thanks for all your answers. We will try the Ribo-Zero kit.

JakobHedegaard
12-13-2011, 07:56 AM
We have done Ribo-Zero Human rRNA depletion of RNA from human clinical samples followed by ScriptSeq library prep and HiSeq sequencing. But it turned out that app 50 % of the reads were of bacterial rRNA origin.
Anyone observed something similar?
/Jakob

pmiguel
12-13-2011, 09:35 AM
We have done Ribo-Zero Human rRNA depletion of RNA from human clinical samples followed by ScriptSeq library prep and HiSeq sequencing. But it turned out that app 50 % of the reads were of bacterial rRNA origin.
Anyone observed something similar?
/Jakob

How was the RNA from the human clinical sample isolated? Were reusable centrifuge bottles or tubes used? These are a frequent source of contamination if not scrubbed and then acid washed.

What was your yield of depleted RNA that you processed using ScriptSeq? We have seen a case where we had very limited RNA going into library construction -- far below specifications for the kit. When we found that a low percentage of the reads mapped, we investigated further. Turned out to be bacterial. I am pretty sure it came from the library construction kit.

Also, if this is a human clinical sample is it possible the patient had a bacterial infection?

--
Phillip

JakobHedegaard
12-13-2011, 11:06 AM
We are trying hard to locate the source(s) of contamination and have so far been able to amplify 16S rRNA fragments from both DNA and RNA - indicating that the contamination has occured before or during RNA and DNA purification.
We start with 50-500 ng total-RNA and from the bioanalyzer profiles of the depleted RNA we have estimated that ScriptSeq is initiated with "just a few ng's" of RNA - should be fine with ScriptSeq. And we do get excellent human data beside the baterial ones.
Did you observe bacterial contamination when using ScriptSeq? Or was it another library prep kit?
Infection of the patient is of course another option...

Cheers, Jakob

Olaf Blue
12-13-2011, 11:42 AM
Jacob, it would be very easy to tell the potential source of the RNA if we have bacterial rRNA sequence information from the sequencing run...if the patient sample IS contaminated with bacteria prior to sequencing, it can certainly be cleaned up with the RZ Kits as I mentioned earlier.

Olaf

pmiguel
12-13-2011, 12:35 PM
Did you observe bacterial contamination when using ScriptSeq? Or was it another library prep kit?
Infection of the patient is of course another option...

Cheers, Jakob

No, it was a completely different kit, several years ago. We were using input RNA at least 1 order of magnitude below the specification for the kit. The data was still usable. But I was a little disappointed that it was contaminated.

--
Phillip