[HeidelbergScience]

Thank you for the information and congrats for finding a company for commercialisation! Do you know whether the custom sequencing primer does also work when CATS libs are multiplexed with other libs, i.e. when the standard Illumina read 2 primer is competing?

Thanks!
Yes, unfortunately, the standard Illumina primer for read2 will compete with the custom read2 primer and they can not be used together.

[seq198]

Thank you for the information and congrats for finding a company for commercialisation! Do you know whether the custom sequencing primer does also work when CATS libs are multiplexed with other libs, i.e. when the standard Illumina read 2 primer is competing?

[HeidelbergScience]

Maybe I missed it, but is paired-end sequencing possible using CATS? Especially for longer ssDNA fragments?

Yes, the paired-end sequencing is also possible, however, only in the commercially offered CATS kits:

Diagenode is now selling CATS kits for RNA-seq:

https://www.diagenode.com/categories/library-preparation

and they also offer a custom sequencing primer for read2 to be purchased separately:

https://www.diagenode.com/p/CATS-paired-end-sequencing-primer-50ul

CATS DNA-seq kits will be launched later, and will also include read2 custom sequencing primer.

[seq198]

Maybe I missed it, but is paired-end sequencing possible using CATS? Especially for longer ssDNA fragments?

[HeidelbergScience]

Is there a publicly available human RNA-seq data set using CATS?

CATS publication data is now searchable on NCBI's Repository (SRA, Short Read Archive)

Project ID is SRP071638.

Here is the link to the project search results:

http://www.ncbi.nlm.nih.gov/sra/?term=SRP071638

Sample Names are as used in the publication.

[Calixer]

Is there a publicly available human RNA-seq data set using CATS?

[HeidelbergScience]

At any rate, this should go without saying, but just a public service announcement since I've met many people who don't seem to be aware: you can reduce your batch-to-batch variation in oligo effectiveness by measuring the true concentrations of your stocks by spectrophotometry, rather than just assuming the manufacturer correctly measured the amount that's in the tube and your resuspension efficiency was exactly 100%. I often find that my "100 μM" stocks are as little as 90 μM, or sometimes even above 100. But then I dilute accordingly before I aliquot.

Very good point, thanks!

[HeidelbergScience]

This is consistent with my tests on a different template-switching protocol: I found that the LNA actually gets slightly lower yields, in every test I've run, except a much higher background of self-priming artifacts. It probably depends on the sequence of the TSO (the Illumina P1 adapter has a run of three cytosines that might hybridize to the G-overhang), but I wonder whether the extra yield reported in the Smart-seq2 paper is actually due to non-specific strand invasion rather than proper template switching (http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3853366/).

At any rate, this should go without saying, but just a public service announcement since I've met many people who don't seem to be aware: you can reduce your batch-to-batch variation in oligo effectiveness by measuring the true concentrations of your stocks by spectrophotometry, rather than just assuming the manufacturer correctly measured the amount that's in the tube and your resuspension efficiency was exactly 100%. I often find that my "100 μM" stocks are as little as 90 μM, or sometimes even above 100. But then I dilute accordingly before I aliquot.

We have made several furthers tests on LNA TSO (rGrG+G). The efficacy (of generating libraries from 5 pg 22 nt synthetic RNA or DNA) was similar to the rGrGrG. However, rGrG+G provided much higher incidence of "empty" libraries, probably due to higher Tm of the +G and the 3'-terminal dC in RT primer. We did not test it on longer templates though.

While rGrG+G was reported to work better on SMART-seq, we would not recommend it for CATS due to its high costs and by-products problem.

[HeidelbergScience]

Can this be modified to work on the Ion Proton platform?

Yes, the CATS can be applied for any NGS platform, including Ion Torrent.
However, a platform-specific sequences has to be used.
Unfortunately we can not uncover Ion Torrent-specific primers here before our next publication is accepted.
However, one can easily derive whose sequences from the structure of the standard Ion Torrent library.

[HappyMonocyte]

Can this be modified to work on the Ion Proton platform?

[jwfoley]

However, if you would read our papers (properly this time) you might notice that this is only one of the changes we introduced in the original Clontech protocol. The effect of betaine and the extra MgCl2 (and, to a lesser extent, the dNTPs in the lysis buffer or the extra cycling at the end of the RT) are as important for increasing the yield.

I did see the parts about the betaine (trimethylglycine), magnesium, and denaturation/activation of the dNTPs, and I reproduced those results successfully. My comment was only about the LNA TSO.

You might also notice (reading the Supplementary), that a TSO with 2 or 3 LNA-G gives actually LOWER yield than the one with only 1 LNA. How do you explain that?

Multiple LNAs in the same oligo aren't typically used in several consecutive positions, especially at 3′ ends, because that makes the molecule too rigid and/or it anneals too tightly to be extended by a polymerase (see e.g. http://link.springer.com/chapter/10....387-32956-0_13). If you had asked Exiqon to design the oligo for you, I suspect they would not have given you those. This is also consistent with the self-hybridization hypothesis.

And I haven´t seen more self-priming artifacts with a LNA-TSO.

As I said, this should depend on the rest of the TSO's sequence. The Smart-seq2 TSO doesn't have more than two C's in a row for the G-overhang to stick to. The Illumina P1 adapter contains three C's in a row, though CATS cleverly truncates the primer to remove those. My test was on a version that included the three C's.

So to clarify, my hypothesis is that the LNA may or may not increase template-switching, but it definitely promotes strand invasion (as shown in the BMC Genomics paper), and it likely creates a problem with self-hybridization if your TSO has a certain sequence. The second factor may account for some or all of the apparently increased yield in Smart-seq2 (which is arguably still a good thing if your goal is to maximize sensitivity), and if it accounts for all of it, that would explain why we don't see the same effect with protocols like CATS that use shorter templates (shorter strands have fewer places to invade instead of annealing at the end). TSOs with more than one LNA don't work for the same reason that other primers with multiple LNAs at the 3′ end don't work.

[Simone78]

This is consistent with my tests on a different template-switching protocol: I found that the LNA actually gets slightly lower yields, in every test I've run, except a much higher background of self-priming artifacts. It probably depends on the sequence of the TSO (the Illumina P1 adapter has a run of three cytosines that might hybridize to the G-overhang), but I wonder whether the extra yield reported in the Smart-seq2 paper is actually due to non-specific strand invasion rather than proper template switching (http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3853366/).

The strand-invasion is a possibility and the risk of that is higher when you use LNA-based TSO. However, if you would read our papers (properly this time) you might notice that this is only one of the changes we introduced in the original Clontech protocol. The effect of betaine and the extra MgCl2 (and, to a lesser extent, the dNTPs in the lysis buffer or the extra cycling at the end of the RT) are as important for increasing the yield.
You might also notice (reading the Supplementary), that a TSO with 2 or 3 LNA-G gives actually LOWER yield than the one with only 1 LNA. How do you explain that?
And I haven´t seen more self-priming artifacts with a LNA-TSO. Or lower yield, for that matter. Obviously it can happen...

[jwfoley]

In our tests rGrG+G and rGrGrG ending TS oligos had equal efficacy (3 independent experiments on 22nt synthetic RNA and DNA).

This is consistent with my tests on a different template-switching protocol: I found that the LNA actually gets slightly lower yields, in every test I've run, except a much higher background of self-priming artifacts. It probably depends on the sequence of the TSO (the Illumina P1 adapter has a run of three cytosines that might hybridize to the G-overhang), but I wonder whether the extra yield reported in the Smart-seq2 paper is actually due to non-specific strand invasion rather than proper template switching (http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3853366/).

At any rate, this should go without saying, but just a public service announcement since I've met many people who don't seem to be aware: you can reduce your batch-to-batch variation in oligo effectiveness by measuring the true concentrations of your stocks by spectrophotometry, rather than just assuming the manufacturer correctly measured the amount that's in the tube and your resuspension efficiency was exactly 100%. I often find that my "100 μM" stocks are as little as 90 μM, or sometimes even above 100. But then I dilute accordingly before I aliquot.

[HeidelbergScience]

I can´t remember this detail off the top of my head, but I would say just slightly worse (a noticeable difference but not huge). All comparisons (Smarter2 vs the Smarter kit) can be found in the Suppl material of the Nat Methods paper.
Suppl table 4 has a list of all the comparisons with details in what they are different between each other.
Suppl table 3 has the detail of all the single cells analyzed with each protocol (sheet A) as well as the statistical analysis of the different variables (sheet B), where we looked not only at the TSO but also PCR enzyme, additives, etc etc.
Suppl Figure 1 and 2 give you an idea of the difference in the sensitivity and variability of different protocols.
In short, the rGrG+G oligo is better than rGrG+N, which is anyway better than the TSO from the kit.

PROTOCOL UPDATE
In our tests rGrG+G and rGrGrG ending TS oligos had equal efficacy (3 independent experiments on 22nt synthetic RNA and DNA).

Importantly however, the quality of the same TS oligo can vary from batch to batch, and affect the yield of the library to some extent.

[HeidelbergScience]

Why do you use T4 PNK instead of a regular phosphatase (e.g. SAP)? Do you need 5’ phosphates for some reason?

Also, why do you say sonication results in 50% 5’ phosphates, 50% 3’? The literature discussed in this thread says there's a "great preference" for 5’: http://seqanswers.com/forums/showthread.php?t=2759

Have you tried using sonicated DNA without phosphatase treatment?

Other phosphatases may also work well, but we did not test them.

We tested sonicated DNA w/o T4 PNK, and the library yield was somehow lower, but not more than 50%.