custom ssDNA library with strange size distribution of qPCR products

[pyridine]

This will take a little explanation. I am preparing custom libraries using the following procedure, which is very similar to the Illumina TruSeq procedure with some minor modifications (aka transposon insertion sequencing):

1. Shear DNA to average 300-bp on a Covaris
2. End repair and A-tail
3. Adapter ligation using custom-synthesized indexed adapters that have the same sequence as in the TruSeq kit, and with another adapter that is only the fragment of the universal adapter that is complementary to the indexed adapter
4. Gel extract products in 250-500 bp range or so (there was a little issue at my workplace when doing this, in that our TAE buffer was far too dilute resulting in a smeary run, so it is possible there is still some adapter present in the samples)
5. PCR enrich with a biotinylated forward primer that is specific to the end of a transposon to pull out transposon-chromosome junctions, and contains the full universal adapter sequence as 5' overhang to create it, plus a reverse primer that binds at the end of the indexed adapter. I have some evidence that this PCR reaction worked as expected, and there was hardly any product visible in the gel above 500-bp.
6. Gel extract products in 200-400-bp range. Biotinylated products should probably run a little slower than non-biotinylated, but probably not enough to matter here.
7. Affinity capture biotinylated PCR products on streptavidin beads, and elute off the non-biotinylated strand.
8. Quantify the single-stranded library with KAPA qPCR kit.

The qPCR appeared worked well and I have amplification curves with a 1:100 dilution that look like the samples are between 1-5 nM. However, I ran the qPCR products on a gel just to see what they look like, and they appear to be centered in the 600-700 bp region. I am trying to figure out how this is possible - if it is just an artifact of the 35 cycles of PCR or if this is actually related to the size distribution of my library? I need to estimate an average size of the library fragments to be able to accurately quantify, and just have no idea what this is! The qPCR DNA standard is 452-bp and ran exactly where it was supposed to. Thanks for any suggestions.

[Genohub]

Sounds like this is an artifact of over-amplification. I would run some of your library on a gel or bioanalyzer before and after qPCR. It could be that you are seeing a "bubble product" forming because of the high number of cycles. I wouldn't be surprised if your library before qPCR ran at the right size.

[pyridine]

Thanks, I agree with your assessment that it is most likely a PCR artifact.

We have been trying to figure out the best way to run ssDNA on a Bioanalyzer - Agilent is basically saying it's not supported but they suggest using the RNA kit - for which the standards are actually ssDNA and they think the dye will bind better than in the DNA kit, but that we should run our own size standard and have to process the data manually. I'm having an awful hard time finding any commercial ssDNA size standards - it basically seems to not exist. It would be highly informative to just compare the library before and after qPCR but the library is ssDNA and the qPCR product is dsDNA, so I don't know if this will be so simple. Perhaps I could just run different numbers of cycles of qPCR and check on a gel if the smear is moving upward with extreme amplification.

[Genohub]

The RNA chip generally runs ssDNA fairly well. Another approach would be to heat denature your qPCR sample at 95 deg for 5 minutes, snap cool and then run on a RNA chip. If your material runs at a smaller size as compared to before heat denaturation, then you were most likely dealing with a bubble product. It turns out that bubble products are innocuous.

[pyridine]

Samples of the original ssDNA library were run on the pico RNA kit and there the size seemed centered around 500-bp, if the size determination for ssDNA is at all correct with it. The areas under each peak trend with the concentrations determined by qPCR. This size range is better but still seems too large, as it should be centered more around 300-bp from the last gel extraction before affinity capture. If this were to then go straight to clustering with a new concentration estimation based on the average size from the RNA pico kit, is there anything else I should be worried about? I'm not doing this part but would like to be aware of pitfalls, as we will be doing this on our own newly purchased MiSeq in the near future. Thanks again.