Amplicon library titration

[HMorrison]

Has anyone used the KapaBiosystems SYBR® FAST qPCR Kits for pre-emPCR titration? And/or know from experience what the optimal starting point molecule to bead input is? Manual says to titrate, which is not likely to happen for multiplexed libraries going on one run. Or, says 4-6 molecules/bead as starting point, which seems high to me.

Thanks.

[Eric@Kapa]

Hi HMorrison,

I was involved in developing the KapaBiosystems Library Quantification Kits, so perhaps I can clarify some of the issues for you. During Beta-testing, and in on-going feedback from some large genome centers that have successfully implemented these kits in their workflows, we have found that:

i) qPCR using reliable reagents and good technique provides very reliable quantification of library DNA for emPCR and bPCR prior to sequencing.

ii) qPCR, by its nature, counts only "PCR-competent" molecules and it therefore provides a more relevant/appropriate and reliable quantification for emPCR (or bPCR in the case of Illumina). However, for this reason (amongst others), there are often significant discrepancies between quant values obtained using various techniques such as bioanalyzer, ribogreen, and qPCR. Generally, we have found that qPCR yields lower quant values compared to techniques that quantify total DNA. This would imply that one might need to add less DNA (measured as copies/bead) when quantification is by qPCR.

Because the kits are still new in the market, we do not yet have sufficient feedback to make general recommendations for copies/bead. Once implemented, our qPCR quantification kits provide users with reliable quant results, and the majority of users have done away with emPCR and/or sequencing titrations altogether. However, optimal copies/bead seem to vary from lab to lab and sometimes from project to project. For these reasons, we suggest that some trial and error and/or titrations may be required initially, in order to determine empirically the optimal copies/bead as determined using our qPCR kits, in your hands.

Which manual are you referring to in your post? To my knowledge we do not make any specific recommendations for optimal copies/bead, for the reasons described above.

I hope this helps; please don't hesitate to contact me for further information, or if you would like to try a sample kit.

Regards,

Eric

[TonyBrooks]

Amplicon library titration

Erik,
The 4-6 molecules per bead is actually given in the Roche emPCR manual (GSFLXTitanium_emPCR_Lib-L_SV_MethodManual_Version2.pdf), page 4.

I'm interested in knowing how people go about quantifying their amplicon libraries for emPCR. We have multiple MID tracked PCR pools. Each amplicon in the pool is the same length. We are torn between using the Kapa kit or just running each pool out on a Bioanalyser 1000 chip to get a peak size and quantitation, then use those to estimate library concentration (MW of dsDNA = 607).

We'll probably emPCR titrate the first few samples to check for reproducibility.
We have about 30 of these pools to run for this project, so titrating every sample isn't very cost effective or time efficient.

Did you find a solution, HMorrison?

[HMorrison]

Well, we found by experience that 4-6 molecules per bead is far too much. The 0.5-2 cpb range is a better starting point for our amplicons. We are trying out the Kapa kit in the meantime. We check amplicon libraries on bioanalyzer and Picogreen each before making a pool, and the RA who does the work likes to pool according to (her subjective opinion of) whether the amplification was strong or weak--puts like products into the same pool before starting emPCR step. She quantitates the pool, then aims for 1+ cpb.

Agreed, titrating each sample is not an option.

[TonyBrooks]

Thanks for that. We know from shotgun we're running at 1cpb, so I guess we might be something similar for amplicon.
Let me know how the Kapa kit looks.

[H5N1]

Well, we found by experience that 4-6 molecules per bead is far too much. The 0.5-2 cpb range is a better starting point for our amplicons. We are trying out the Kapa kit in the meantime. We check amplicon libraries on bioanalyzer and Picogreen each before making a pool, and the RA who does the work likes to pool according to (her subjective opinion of) whether the amplification was strong or weak--puts like products into the same pool before starting emPCR step. She quantitates the pool, then aims for 1+ cpb.

Agreed, titrating each sample is not an option.

Agree! From my recent amplicon runs, 0.5-2 cpb gave good results. At 2 cpb, the enrichment% is around 20%, even at 20%, still produce very good reads.

But it also depends on the library. I had some amplicons requiring 8 to 16 cpb to produce 8-10% enriched beads. However, the read quality of these beads is not as good as others. They often have higher portion of short reads.

The newer version of manual mentioned 4-6 cpb, which seems too high. But it also emphasizes the use of siliconized tube for bead enrichment. I haven't got the siliconized tube. I wonder if this will reduce the enriched percentage and is the reason why require higher DNA input?

[AllIn1ng]

Hello everybody,

is it possible do determine the average fragment length of libraries for 454 titanium sequencing even if the concentration of your sample is too low to detect it with the bioanalyzer, since you need that data for calculating libraryconcentration? Maybe one could clean up the KAPA-qPCR product to analyze it on the hs dna 7500 chip. Or does the PCR change fragment length composition of our Libraries?

[TonyBrooks]

Hello everybody,

is it possible do determine the average fragment length of libraries for 454 titanium sequencing even if the concentration of your sample is too low to detect it with the bioanalyzer, since you need that data for calculating libraryconcentration? Maybe one could clean up the KAPA-qPCR product to analyze it on the hs dna 7500 chip. Or does the PCR change fragment length composition of our Libraries?

Can you not run a library where you know the optimum cpb (by emCPR) along with your new test sample. Just ignore the cpb calculation based on read length.

[Eric@Kapa]

Hi AllIn1ng,

Which 454 library prep protocol/workflow are you following? If you are doing a size selection (gel purification or SPRI bead) at some point, then perhaps you can use that size range to infer the likely fragment sizes in your library. Alternatively, it should be fine to size the qPCR products on a gel or bioanalyzer, but bear in mind the following:

i) If amplification occurs relatively early (e.g. around 10 cycles), and the qPCR protocol continues for many more subsequent cycles (e.g. 35 cycles, so an additional 25 after crossing the Ct in the e.g. above), then the final PCR products are likely to have both large- and small-sized artifacts that are not representative of the actual library composition (e.g. "primer dimer" type products, and large "concatemerized" and/or partially single-stranded products).

Depending on how variable your library concentrations are, you might be able to minimize these effects by cutting down your qPCR cycling protocol to the minimal number of cycles to produce the standard curve and to obtain Cts for your samples (e.g. ~20 cycles).

ii) Any SYBR Green carried over from the qPCR will significantly affect the mobility of the qPCR products in a standard agarose gel. I'm not sure on the effect of SYBR Green on a Bioanalyzer run. Although a PCR-cleanup column procedure might slightly affect the size distribution of an (amplified) library sample, it shouldn't be dramatic in the relevant range if you use a good product from a reputable supplier (e.g. Qiagen?). You might also experiment with running samples of the amplified qPCR DNA standard(s) along with your samples to help account for likely artifacts (though these might differ between actual library samples and the DNA standards), and any effects from SYBR Green dye.

iii) Extremes aside, the adjustment calculation made for fragment size shouldn't make a dramatic difference to the final answer, since we tried to use DNA standards for our 454 library quantification kits that are "typical" or "average" (around 450 bp). In most cases we would expect libraries to comprise fragments around this size range, so the adjustment is not likely to change quant results by more than ~20 - 50% in either direction. Assuming you make a reasonable "educated guess" about the fragment sizes in your library, this factor should be further reduced.

[Eric@Kapa]

Hello everybody,

is it possible do determine the average fragment length of libraries for 454 titanium sequencing even if the concentration of your sample is too low to detect it with the bioanalyzer, since you need that data for calculating libraryconcentration? Maybe one could clean up the KAPA-qPCR product to analyze it on the hs dna 7500 chip. Or does the PCR change fragment length composition of our Libraries?

Hi AllIn1ng,

Sorry for the late response. There are a few options in cases like this:

i) Do you have good information about the size distribution of your starting material (e.g. sheared gDNA)? If so, this might act as a guide..?

ii) If you are performing some sort of size selection after adaptor ligation (e.g. gel purification, Pippin Prep, double-SPRI), then this might act as an indication of the expected size range of fragments in your final library, even if you can't actually see the material on a Bioanalyzer.

iii) Your suggestion of analysing the qPCR products is viable, but you are correct to be cautious about changes during PCR. We have run comparisons of our engineered (evolved) qPCR polymerase against various wild-type competitor qPCR reagents, and one definitely sees a significant shift in size distribution (towards smaller fragment sizes) after amplification with the wild-type products. The good news is that we didn't see any significant size bias on a library of ~ 200bp - 800bp fragments after amplification with KAPA SYBR FAST qPCR reagent.

However, even if you are using the KAPA Library Quantification Kit, there may be some changes in size distribution during qPCR. The greatest concern in this respect is that samples are generally relatively concentrated, having Ct scores <10. If you are running a "standard" qPCR protocol, then these samples might be undergoing >20 cycles of PCR after they have crossed the Ct, and during this phase dNTPs and/or primers are likely to be exhausted, leading to a variety of artifacts comprising both low- and high-molecular weight species (e.g. "primer dimer" type effects, as well as incompletely extended ssDNA, concatemers, etc.). Thus, it would be best to try to reduce the number of qPCR cycles as much as possible, and to treat any peaks outside of the the central "main" library smear with suspicion.