Unexplained short reads on PGM

[greggiceton]

Hello all. I wonder if anyone has experienced this, and if so how you resolved the problem?

Using our OneTouch V1 with the 200bp V2 DL kit and PGM with 200bp sequencing kit I am experiencing a large peak ~45bp from a 16S V3 library that should be closer to 200bp after barcode and adaptor trimming. Both instruments are working correctly. Our Qiaxcel Advanced shows the short fragments are not "real". Pre-enrichment figures are good (20%). I spiked in the 200bp control library one the same chip and got a mean read length of 216bp. The only cause that remains on the table is the amplicons forming hairpins or some such, which makes no sense as a) lots of other people use V3, b) another group has achieved good results with the exact same protocols, and c) I would expect my library prep to fail were the amplicons forming hairpins. I am using fusion primers with a 10bp barcode and have confirmed that the primers do not form hairpins. Any insight would be much appreciated as this is driving me crazy!

[aggp11]

Hi,

Could you share a snapshot of the read length distribution plot that is generated as part of the report? Also, what is the size of the library you are using for sequencing after adding the adapters and barcodes?

[snetmcom]

Do you have the image for your Qiaxel? Does it have the resolution for pg amounts of adapter dimer?

You could also look at the most populated short reads. That would tell you for sure.

[greggiceton]

Hi. Read length histogram attached - this is for the entire run. The control library histogram (tiny little thing that can't be expanded annoyingly) has no short peak. The amplicons *should* be 258bp inc adaptors and barcode.

Resolution for the Qiaxcel cartridge in use is 20-50bp so not ideal, but it detected nothing shorter than 269 with a marker detected at 15bp.

However having been given a nudge in the right direction I have discovered that the short reads are a combination of the forward primer (after removal of adaptor + barcode) and a fragment of the reverse compliment of the reverse primer (after removal of the adaptor). Analysis on Oligo Analyzer confirms the fragment matches the region of heterodimer overlap at -16 dG, so it seems I have lots and lots of heterodimer. Which begs the question why did the Qiaxcel not detect them and how did these 69bp (inc adaptors + barcode) fragments get through eGel size selection and then AMPure XP clean up which is supposed to select for fragments >100bp?!

Next step is a comparison of fewer cycles, multiple clean ups and touch down PCR. Lab work is hard!

[msuokas]

I think general concensus is that fusion primer dimers are able to anneal with the amplicon molecules. We certainly have seen this in several amplicon samples. This seems to happen most often in samples that are amplified with universal primers.

Only solution we have found for problematic samples is to purify using ampure XP as many rounds as necessary (often 6-9 rounds). We will check the 454 samples with pcr using specific A- and B-primers (~20nt long) and Ion Torrent samples with A- and trP1-primers. If you have trace amounts dimer molecules in your sample, they will be ampilified.

[JOM]

We had same problem and we also tried several ampure rounds, but what definitively works for us was load the library on an agarose gel, select the right size and gel extraction.

[greggiceton]

Hi. I ran 4 samples recently comparing single and double AMPure cleanup and double cleanup followed by EGel size selection on two different DNA extractions. The results are substantially better from initial analysis (mean read lengths of 185 and 160bp) with the most significant peak at 200bp as expected. I will try 3 x cleanup with and without EGel next I think. As for cutting out the band I have heard of others having success with this - I just can't understand why that works yet EGel doesn't!

[snetmcom]

have you tried adding less primer or performing less cycles? You don't need much library to run.

[greggiceton]

The last of my 5 previous PCR runs was with less primer so that will be one of my next comparisons. I will also look at less cycles as I progress through the various options.