I wouldn't recommend using Nextera to do RNA-Seq as you lose strand specificity and the kit itself is very expensive. We use the technique from the paper below and it costs us around $5 to create a sequencing library from total or rRNA depleted RNA. There are also other methods available that can produce strand-specific libraries at a lower cost than using Nextera. If you already purchased all of the reagents you mentioned, what you suggested will work relatively well, except you will lose strand-specificity and it will likely take a day or two to complete the process.

http://www.ncbi.nlm.nih.gov/pubmed/23558773

I wouldn't recommend using Nextera to do RNA-Seq as you lose strand specificity and the kit itself is very expensive. We use the technique from the paper below and it costs us around $5 to create a sequencing library from total or rRNA depleted RNA. There are also other methods available that can produce strand-specific libraries at a lower cost than using Nextera. If you already purchased all of the reagents you mentioned, what you suggested will work relatively well, except you will lose strand-specificity and it will likely take a day or two to complete the process.

http://www.ncbi.nlm.nih.gov/pubmed/23558773

I have not bought the material yet.... where could I get the Peregrine primers?

Taking a look at whole material for what you suggest, it does not seem to me that there will be a significant decrease in price.... I still will have to by both kits for ds cDNA synthesis and also the Epicentre kit for barcoding.... am I wrong?

Also, one more question: the synthesis of ds cDNA starting from random primers will give me a mixture of fragment sizes. For what I have been told Nextera is ok if fragment size is > 300 bp.

Will I loose information in this step?

Unless you require sequencing data for rRNA you will need to use a depletion kit such as RiboZero from epicentre (http://www.epibio.com/applications/rna-sequencing/rrna-removal/ribo-zero-rrna-removal-kits-(bacteria)).

6 reactions costs around £300 but I routinely use this kit to prep bacterial transcriptome libraries and you can half the volumes of the reagents used (just normalise your total RNA in 14uL of water not 28) this method will take around 3 hours and removes the vast majority of ribosomal material.

After enrichment we use the NEB next Ultra directional kit. Although expensive, as you say (around £30 per sample). It does provide directionality and is very easy to use. There are minimal clean up steps, and the whole prep (from enriched RNA) can be accomplished in one day.

Another question about using random primers to synthesize my cDNA:

would random primers bind more than once to my transcripts giving me a bias in transcript representation in the total population?

If you use the NEB next kit the random primers are added during fragmentation of the RNA removing this bias.

If you decide to use a custom kit then I would expect that any bias generated by creating a higher number of products from longer transcripts would be out weighed by preferential amplification of smaller species? There may be a bias towards the extreme ends however (i.e smallest and longest transcripts).

I would think that in the end if you sequence deep enough the bias shouldn't be a problem as it could be removed computationally.

Thanks for these answers! I think that my problem is that I am trying to decrease the costs so I can`t use the NEB next kit. Moreover, I am starting from prokaryotic RNA that does not contain polyA. So, my fera is that two of the following options will happen :

1) if fragments are too small I use information...

2) if more primers bind to the same longer RNA, this will then be more represented in the sequencing introducing a bias...

I do not know if I can fix this in some way...

The ribozero kit contains specific probes for rRNA which you anneal and effectively remove from solution leaving all other RNA. This means you don't have to worry about the issue of prokaryotic RNA being polyA - . Without using this kit ~96% of the sequencing data generated will correspond to rRNA. Unless you are studying ribosomal RNA I would say this kit is essential.

1) Losing fragments should not be a problem in any prep that I'm aware of. If you really want to maintain all RNA including small RNA fragments etc, you can adjust the bead cleanup to a ratio of 3:1 beads to RNA this should ensure minimal loss of any material. I could not be sure how this would affect subsequent stages in the prerp and at a minimum may reduce yield. In all protocols custom or kit there will be a size selection step where material is lost but as long as your size selection in broad you should have no problem.

2) As far as I'm aware the huge amount of data generated during sequencing will mean that any slight bias such as primer binding or loss of smaller fragments can be accounted for bioinformatically by sequencing depth. Essentially your libraries will contain RNA representing the whole transcriptome. However I would be concerned that without using a robust kit you may have issues of even coverage etc and could cause problems further down the line in terms of analysis.

As a further note by using the nextera kit you will lose 3' and 5' data as the tagmentation protocol cannot produce libraries from these areas due to the cut and tag nature of the transposon, requiring a sequence on either side to bind to. This method is fast and simple but comes at the cost of reduced coverage in some areas.

I hope this is some help!

I agree with Martin that you will have to do something to get rid of the rRNA to get an informative library. We have found that in E. coli, rRNA makes up ~95% of the total transcripts. While Ribo-Zero is an excellent option to do this, it's pretty expensive. You may want to try DSN treatment. It's an enzyme based method that adds a day to your prep time, but costs a fraction of the Ribo-Zero kit. If you were to use the peregrine method there would be some start up costs with ordering primers, enzymes, buffers, etc, but when you get these items you will be able to do hundreds of reactions rather than the 24 reactions that Nextera allows. It's the bulk ordering the saves you money and leads to the $5/sample cost that I mentioned previously. We order our primers from IDT and everything else from Clontech. Incidentally, about 6 months after we published our paper, Clontech came out with a kit that does the identical reaction, but costs ~$100/sample. As for bias using random hexamers during cDNA synthesis, you will see a drop in coverage at the 3' end of the transcript, but as Martin said, using Nextera you will see the same thing plus a drop at the 5' end as well.

As Zachary and Martin pointed out, getting rid of the rRNA is essential and the RiboZero kit seems to be the best way to do this. You might try Terminator exonuclease but the secondary structure of the stable rRNA will be a problem. We worked with Bacillus and a number of high G+C Actinobacteria so far, and RiboZero did by far the best job in all instances.

Concerning synthesis of the cDNA, I would not use the Nextera kit. From my experiences, losing directionality and information on the 5'- and 3'ends is not a good thing. If you just want to do DGE that might be tolerable. For everything else, like detection of TSS, novel transcripts, or operons, you are likely to create a nightmare for data analysis. I got data sets from different collaborators and quite often there was no way to get out the information they were looking for.

Beside the protocol Zachary mentioned, you might take a look at the ones we developed for total transcriptome end 5'-enriched libraries

http://www.biomedcentral.com/1471-2164/14/888

and the protocol for small RNA capture and sequencing.

http://www.biomedcentral.com/1471-2164/14/714

With those, we got a really good resolution of TSS and operons and found quite a number of novel transcripts. Alas, the protocols are quite time consuming and some components are not cheap. But you end up with ready-to-sequence barcoded libraries. If you are interested, I can sent you the current version of the protocols, as optimization continues.

I am also preparing to do RNA-seq, but on plant transcriptome. We have well over a thousand samples, so I need a cheap library prep. The method Zachary mentioned is supposed to cost $5/sample, but looking at the paper I noticed they're using 20-nt tags, which I understand will be sequenced together with the transcript. Even if I'm doing 100 bp reads, 20% of the sequence information will be lost to the tag, making me have to sequence more and driving up the cost. Can you comment on this, Zachary?

Hi Justyna,

We use a custom sequencing primer that includes the added tag. This way the first base you see is part of the transcript rather than the tag. The primer is called PP_R1, and its sequence is in the methods section of the paper on page 512. Incidentally, if we didn't use a custom sequencing primer the run would fail because the first 20 bases of every single read would be identical and the optics wold be unable to differentiate clusters.

Zach

That makes total sense now. Thank you, Zach. And I apologize for not reading your paper close enough; it was mentioned there.

I think this method looks very promising for our needs. I'll definitely look closer into it.