We would like to make RNA-seq libraries from Drosophila RNA. It is important that we deplete rRNA rather than enrich for Poly(A) as we would like to analyse some non-polyadenylated species in our datasets.
I have read that the Illumina Ribo-Zero rRNA depletion kit (Human, Mouse, Rat) is reasonably effective, although does not fully remove 5S or 28S (right arm) rRNA.
I'm also thinking about an rRNA "RT blocking" rather than "depletion" strategy a la the method presented in the attached paper, although I would have to devise a strategy for removing all types of rRNA (not just 2S).
Has anyone had experience with this, and would you please share some wisdom? What sort of parameters would I have to be careful to consider using these approaches? Do you have alternative suggestions?
As an aside, these RNA-seq libraries are likely to be from low-input RNA, and the quality could be less-than-perfect as we are isolating RNA after formaldehyde fixation.
Thank you in advance for your intelligent answers.
Hi all,
Apologies for the radio silence!
It turns out that the homebrew biotinylated probe depletion works a treat (with some caveats).
The steps:
1) Design PCR primers to amplify rDNA genes from genomic DNA. You want to cover basically the whole RNA sequence. For longer species - 18S and 28S (L/R) - we amplified two halves of each gene (so, two probes for 18S and a total of four probes for 28S)
2) In vitro transcribe your rDNA PCR products by incorporating Biotin-UTP into the mix. There are plenty of different ways to do this - choose your favourite
3) Quantify your RNA probe concentration using Qubit and pool your probes so that there is at least a 5x excess of probe:RNA (this requires optimisation, sorry)
4) Hybridise your probe mix to RNA sample, followed by incubation with strepavidin-coated magnetic beads.
5) Keep supernatant! This is your rRNA-depleted sample.
Each step requires optimisation. I found that the IVT (despite being quite a routine procedure) was quite temperamental. You can assess depletion using qPCR against rRNA species - I find this is much more quantitative and informative than Bioanalyzer analysis.
NOTE: you may have to tinker with the probe pooling concentrations. I found that if I add too much excess that the beads don't completely remove the probe and then when you do the qPCR it looks like the rRNA levels have actually increased!
CAVEAT: I had success removing all species but 5S. This is the same as Ribo-Zero. Not something I would lose sleep over as 5S is smaller than most library prep size cutoffs (usually 200nt+). Maybe you will have more success?
However, after all of this, I have all-but-abandoned this protocol for RNA-Seq. We are currently using a kit from NuGen (Ovation RNA-Seq System for Drosophila - http://www.nugen.com/products/ngs/ovation-rna-seq-systems-1-16-model-organisms) that is VERY reasonably priced per reaction and has a nice "rRNA exclusion" step where - rather than depleting - it excludes unwanted RNAs from PCR amplification in the final step. It's very nice, and in our first samples we saw between 1.5-5% rRNA in our final sequencing libraries. Plus, you can customise it to exclude other contaminating RNAs.
If anyone has questions, feel free to shoot me an email at [email protected]
Thank you for your answers. Andrei, I have followed your advice and am in the process of designing complementary probes to Drosophila rRNA for RNase H depletion. I will post my results using this approach.
Oligo-dT columns would allow you to deplete the rRNA. You would have to make some controls to see how much rRNA do you retain in your sample after the column, and how much mRNA do you loose in the column
Best
Jose A.
@Olga, I have been in contact with NuGen about this kit. Unfortunately, the rRNA depletion part of the kit is not available alone (ie. you have to buy the whole NGS library prep set). This is no good for us as we have our own kit-free library prep method. But for you, it is definitely worth testing. If you ask nicely, they will give you a 50% discount on the first "test" kit, for 32 samples. Good luck!
Andrei,
I am going to test a homebrew probe depletion strategy. The homebrew probe depletion involves several steps:
1) Design PCR primers to amplify the entire rDNA region for each rRNA subtype, and use regular genomic DNA as your template. The reverse primer must have a T7 (or similar) promoter sequence at the 5' end (TAATACGACTCACTATAGGG).
2) Perform in vitro transcription with Biotin-UTP using T7 polymerase to transcribe antisense RNA probes. Digest the DNA with DNase and clean up RNA with phenol or column clean up.
3) Hybridise these antisense probes to the total RNA sample (containing rRNA) and sequester the rRNA:probe complexes with Streptavidin beads
4) Check rRNA depletion using Agilent Bioanalyser, qPCR against rRNA genes, and in the final instance, using RNA-seq.
I am in the process of testing this approach with Ribo-Zero - will keep you updated when I have some results.
:)
Hi all,
Apologies for the radio silence!
It turns out that the homebrew biotinylated probe depletion works a treat (with some caveats).
The steps:
1) Design PCR primers to amplify rDNA genes from genomic DNA. You want to cover basically the whole RNA sequence. For longer species - 18S and 28S (L/R) - we amplified two halves of each gene (so, two probes for 18S and a total of four probes for 28S)
2) In vitro transcribe your rDNA PCR products by incorporating Biotin-UTP into the mix. There are plenty of different ways to do this - choose your favourite
3) Quantify your RNA probe concentration using Qubit and pool your probes so that there is at least a 5x excess of probe:RNA (this requires optimisation, sorry)
4) Hybridise your probe mix to RNA sample, followed by incubation with strepavidin-coated magnetic beads.
5) Keep supernatant! This is your rRNA-depleted sample.
Each step requires optimisation. I found that the IVT (despite being quite a routine procedure) was quite temperamental. You can assess depletion using qPCR against rRNA species - I find this is much more quantitative and informative than Bioanalyzer analysis.
NOTE: you may have to tinker with the probe pooling concentrations. I found that if I add too much excess that the beads don't completely remove the probe and then when you do the qPCR it looks like the rRNA levels have actually increased!
CAVEAT: I had success removing all species but 5S. This is the same as Ribo-Zero. Not something I would lose sleep over as 5S is smaller than most library prep size cutoffs (usually 200nt+). Maybe you will have more success?
However, after all of this, I have all-but-abandoned this protocol for RNA-Seq. We are currently using a kit from NuGen (Ovation RNA-Seq System for Drosophila - http://www.nugen.com/products/ngs/ovation-rna-seq-systems-1-16-model-organisms) that is VERY reasonably priced per reaction and has a nice "rRNA exclusion" step where - rather than depleting - it excludes unwanted RNAs from PCR amplification in the final step. It's very nice, and in our first samples we saw between 1.5-5% rRNA in our final sequencing libraries. Plus, you can customise it to exclude other contaminating RNAs.
If anyone has questions, feel free to shoot me an email at [email protected]
Here is a well cited paper, that worked for me and others in the field.
http://www.silencejournal.com/content/3/1/9
Hi Nikolay,
Thank you for the paper. I think the key to this discussion though is to devise a protocol that does not use commercial depletion kits such as Ribo-Zero, as this is an expensive approach when preparing many individual libraries for RNA-seq.
Ali
- Badges
- Science topic
- Similar topics
- Biological Science
- Cell Biology