primer design for low diversity libraries

[btb]

Hi. We are using a transposon sequencing (Tn-seq) protocol that uses a custom sequencing primer to sequence through the transposon into the genomic DNA of interest. The lack of diversity in the transposon sequence is causing problems on HiSeq runs. We are considering changing the protocol to incorporate something like the Illumina universal adapter during library prep and I have some questions about how to resolve the low diversity issue (as we would still like to sequence through the transposon). I know that staggered primers can be used to introduce diversity. My question is whether using a single primer with a 'variable' region (6-8 N's introduced during oligo synthesis) would also work? In this case, one would sequence through the variable region, then into the low diversity region, then into the high diversity region of interest. Use of a single primer would be more cost effective than staggered primers. Does anyone have any experience or thoughts on the matter?

Thanks.

[HESmith]

Yes, this approach will work, but you'll want to use two rounds of PCR (first, amplicon primers; second, adapter-variable-amplicon primers). If you use single-round PCR with the variable region primer, you can introduce bias in the variable sequence via extended base-pairing with the template that flanks the amplicon target sequence.

Alternatively, you can spike in phiX control or a high-diversity library with your existing samples. The recommended amount is specific to the sequencing instrument.

[btb]

Thanks for the response @HESmith! The protocol does in fact use two rounds of PCR. We use a Terminal transferase to add a poly-C tail to the sheared and end-repaired DNA. The first round of PCR uses a (forward) Tn-specific primer and a (reverse) primer that binds to the poly-C tail. Round 2 PCR uses a nested Tn-specific primer that would include the Illumina adapter and variable region sequence, while the reverse primer binds to the sequence added by reverse primer in round 1 PCR. If I understand correctly, the nested nature of the round 2 forward PCR primer could still introduce bias in the variable sequence via extended base-pairing with the template that flanks the amplicon target sequence. Perhaps we should change the strategy to avoid a nested PCR?

We are trying to avoid high-diversity library/phiX spike-ins to retain reads, but we'd also like to find a protocol that makes our libraries compatible with any sequencing chemistry/platform to reduce headaches and increase flexibility in lane sharing.

Do you know of any literature that uses a variable primer like we've discussed? Also do you have any thoughts on how long the variable region needs to be? I was thinking 6-8 N's.

[HESmith]

Not sure what you mean by nested PCR. That usually implies a primer that's interior to the template. The design I mentioned would more accurately be described as an overhanging primer (adapter-variable-Tn, where Tn corresponds to the 5' end of the original amplicon). In this design, there's no template for the variable bases to pair with.

This design is the same as random barcodes for discriminating independent library molecules. You should be able to find relevant publications without too much effort.

[btb]

You are correct-- our round2 forward PCR binds interior to the round1 forward PCR primer, so there would be template for the variable bases to pair with. We could easily redesign the scheme to instead use an overhanging primer.

Thanks again, HESmith!