Clontech Stranded mRNA-Seq for Illumina: Two peaks in library?

[Innovelty]

Hi folks,

Application: poly-A RNA-Seq on a MiSeq, 2x250 pe reads.

Our desired library insert size is between 500 and 800 bp, though of course getting a size distribution that tight is tough. So far I'm not anywhere near. We've gone back and forth with Clontech several times about how to "titrate" the Mg++ fragmentation times to achieve less fragmentation, but they've been largely unhelpful. I finally decided to use a fragmentation time of 2:30 at 94C, on the poly-A-selected RNA represented by the Bioanalyzer traces in the file "mRNA_input.pdf" attached.

Other relevant details:
- 50 ng input RNA
- 12 cycles of PCR
- used 0.5x volume AMPure beads in each cleanup step, as suggested by Illumina in the TruSeq and Nextera manual -- my calibration of the beads suggests that that should cut everything below 300 bp.

So, what's with these libraries with two peaks at ~150-180 and at ~300-400?

I should say that I have had nothing but trouble with the Bioanalyzer, and do have the opportunity to size my libraries with the QIAxcel machine. Is it worth running that before I make a decision on these? Should I just redo the libraries and, if so, what should I change?

[MU Core]

Hello Innovelty,

The extra peak around 400bp will be due to over-amplification during PCR. If the primers become limiting the adapters on the ends will anneal between non-identical fragments which result in the observed higher sized molecule. The library will sequence fine.

As for fragmentation of the mRNA, the divalent cation method for shearing RNA is very quick so I don't think it possible to obtain the larger insert sizes you desire other than by first making full-length cDNA and shearing by other method such as a Covaris.

[Innovelty]

Thanks, MU Core, after reading some more here that's what I figure is probably going on, as well. I tried forum user pmiguel's recommendation of heat denaturing and running on an RNA Pico chip (as seen here: http://seqanswers.com/forums/showthread.php?t=12523), but that did not actually look any better, and was a bit more confusing.


So could we sequence these and just go for fewer cycles? And, how can I determine the size of these libraries in order to figure molarity for loading onto the machine?

[MU Core]

You'll want to use the peak size of 187 for determining molarity.

Try using one cycle less in the PCR amplification step on future libraries. This should reduce/eliminate the extra peak and reduce any potential PCR bias.

[nucacidhunter]

And, how can I determine the size of these libraries in order to figure molarity for loading onto the machine?

I would suggest taking a small aliquote of the library and doing a PCR for limited cycle followed by sizing on TapeStation or Bioanlyser. That way you will see the size and it will be a test as well whether overamplification was the cause as you would not expect to see that larger peak this time.

[bilyl]

Hi folks,

This might just kill me, it honestly might.

Application: poly-A RNA-Seq on a MiSeq, 2x250 pe reads.

Our desired library insert size is between 500 and 800 bp, though of course getting a size distribution that tight is tough. So far I'm not anywhere near. We've gone back and forth with Clontech several times about how to "titrate" the Mg++ fragmentation times to achieve less fragmentation, but they've been largely unhelpful. I finally decided to use a fragmentation time of 2:30 at 94C, on the poly-A-selected RNA represented by the Bioanalyzer traces in the file "mRNA_input.pdf" attached.

Other relevant details:
- 50 ng input RNA
- 12 cycles of PCR
- used 0.5x volume AMPure beads in each cleanup step, as suggested by Illumina in the TruSeq and Nextera manual -- my calibration of the beads suggests that that should cut everything below 300 bp.

So, what's with these libraries with two peaks at ~150-180 and at ~300-400?

I should say that I have had nothing but trouble with the Bioanalyzer, and do have the opportunity to size my libraries with the QIAxcel machine. Is it worth running that before I make a decision on these? Should I just redo the libraries and, if so, what should I change?

Why don't you just try running the bioanalyzer on the RNA right after the fragmentation process? That would give you an idea of what to expect. You could also do it on 1ul of the sample right after the second strand synthesis cleanup with a high sensitivity DNA chip and that will for sure let you see what is happening to the size distribution.

Also, the KAPA mRNA kit recommends 85C for 6 minutes to get 300-400bp. Maybe you can try that temperature for ~4-5 minutes to get it up to ~500?

[Clontech]

Many thanks to Innovelty for following up with our Technical Service Representatives and providing additional information and experimental details.

After reviewing the data, we concluded the following:
- The peak of the fragmented RNA coincides with 18S rRNA, which usually runs at about 2,000 nt (refer to http://www.nature.com/nmeth/journal/....f.364_F1.html)

- The sheared RNA was long enough to generate cDNA and the cDNA was retained by SPRI bead purification. (Note: the cumulative size of SMARTer® Stranded primers is 125 nt.)

- We also noticed that the ds cDNA profiles are highly unusual. We think that 180 and 400 bp peaks of Sample 84 are most likely artifacts (samples 87 and 50 show SPRI bead contamination). We recommend running a negative control, as doing so enables data evaluation.

- It is possible that ds cDNA was lost during SPRI bead purification.

We suggest taking the following steps to troubleshoot:
• Try a 0.7:1 SPRI bead: sample ratio to purify first-strand cDNA. Our testing indicates that 0.8:1 ratio correspond to a peak at ~350-400 bp, and 0.6:1 corresponds to a peak at ~650 bp, so 0.7:1 may be most suitable for your purpose.
• Add more PCR cycles to generate more ds cDNA.
• Include a 1:1 final bead clean-up after PCR.
• Again, be sure to include a negative control.
• Based on the attached PDF, “Heat Degradation Time Course for Full-Length poly(A) RNA,” you may consider evaluating heat degradation as a means of RNA shearing, with a 2 min incubation step. Lowering the temperature for fragmentation may help provide control over the fragmentation process, as it occurs very rapidly. Alternatively, Covaris might be considered.
• Proceed with the protocol of the SMARTer Stranded RNA-Seq Kit User Manual, taking care to follow all steps as written.

Please do not hesitate to follow up with our Technical Support Representatives at [email protected] or 1.800.662.2566.

[Innovelty]

These problems are solved. I'll attach a bioanalyzer readout. Here are the relevant details regarding how I fixed it, and many thanks to Clontech Tech Support for holding my hand through it.

Input mRNA: 20 ng
Fragmentation: 2 minutes at 94 C then snap-cooled and held for at least 2 minutes on ice
SPRI ratio for cDNA cleanup: 0.7x volume AMPure beads to sample (0.7:1)
PCR cycles: 11
SPRI ratio for PCR cleanup: 1:1



The "ML" library is a positive control with included mouse liver RNA, while 2029 is my own sample. The shortest fragments are just about exactly where I want them, and while we have a pretty broad peak, I think it would be hard to get a tighter band without going the Covaris route. These look a lot like my Nextera libraries which sequenced fairly well.