Custom MiSeq sequencing primers

Custom sequencing primers opens the door to a lot of possibilities, particularly when attempting amplicon sequencing, allowing you to sequence off your amplification primers and not waste high-quality read data on the one bit of sequence you already knew.

Here follows a list of considerations one might want to take into account when generating custom sequencing oligonucleotides for use on the Illumina MiSeq machine (ignoring all the low diversity issues inherent in Illumina sequencing).

I write this with full disclosure from the outset; I'm relatively new to NGS, new to the MiSeq, and the only runs I've tried using custom primers have failed abysmally. However, in troubleshooting why these might have failed (to feature in a later, longer post) I've assembled a fair bit of information on custom primer design, which might be useful for others (or others might be able to point out where I've gone wrong!).

First, the basics. The MiSeq is the latest offering from sequencing powerhouse Illumina, providing a benchtop alternative to it's larger previous models. It uses the same Solexa-style chemistry as the Genome Analyzer and the HiSeq, but in a smaller box, aimed at lower-throughput applications.

The important thing for this post, is that the MiSeq does support the use of custom primers. These can either be used independently, or spiked into the primer reservoirs for read one or two (see p73 of the MiSeq System User Guide, access probably requires free sign up).

The question then becomes how to design these primers.

The first consideration practically goes without saying; the primer needs to prime, and only where you want it to. Were there to be more than one complementary site per cluster there would be multiple bases seen per cycle, and the cluster would be thrown out.

Within that constraint, the basic idea typically is then to match the characteristics of the Illumina sequencing primers, the sequences of which can be found in the Illumina Customer Sequence Letter (obligatory copyright notice; oligonucleotide sequences © 2007-2012 Illumina, Inc, all rights reserved, yada yada yada).

(By the way: In assessing the Tm of your oligos, apparently Illumina tech suggest using the IDT calculator)

The Illumina sequencing primer 1 (SP1) is 33bp long, 51.5% GC, with a Tm of 65.5°C, while SP2 is a bit longer/tighter binder, being 37bp long, 59.5% GC and having a Tm of 70.1°C. Note that the MiSeq runs hotter than the HiSeq during the deblocking and extension stages, meaning that the Tm of your oligo has to exceed 65°C in order to prevent dissociation from the target.

Logistically speaking, the most important of these criteria to match (or exceed) is probably the Tm; the length and GC content will largely be dictated by your sequences*.

If the primers used to generate your amplicon have a Tm lower than 65°C, there are several workarounds you can use to bring the melting point up to scratch.

One option is to extend your sequencing oligo into the P5 or P7 element that it borders, akin to how the Illumina SP1 oligo overlaps P5 by 5 bases. A note of caution; Illumina tech have warned me that too long an overlap could result in non-specific annealing of primers to the lawn of adapters. I can't see that this would ever result in mis-calling, as the complex would have a huge 3' overhang (therefore not extend, thus contributing no fluorescence), but it could certainly reduce the effective concentration of sequencing primer in the mix.

Another similar sensible option suggested to me recently by Tony Brooks (of UCL Genomics) is to insert a few dummy nucleotides to the 5' of your PCR primer, and then include these in the sequencing oligo, bringing the Tm up without entering the adapter elements.

Lastly, a powerful but potentially tricky option; use modified bases in your sequencing primer to increase the stability. Particularly, the use of Locked Nucleic Acid (LNA) containing oligonucleotides (the subject of a future post I think) has received a lot of talk as a potential way to boost the Tm/specificity without increasing the length.

Custom oligos open doors to a lot of useful and novel sequencing strategies. However, as I'm finding out - to my continued frustration - they do present a wealth of problems, which by their very nature are unique, and of course, unsupported.

If anyone has any pointers, corrections or other advice, please do share in the comments.

TL:DR: Check this nice explanation out. It has pictures.

* As a little aside, it's worth noting that binding of a primer to its cognate site isn't necessarily the key determining factor of its ability to prime; certain motifs may be preferentially favoured by polymerases. However, this is hard to measure; the only hints I'm aware of are a tantalising blog post and a patent application from prolific professor Jian Han (from the HudsonAlpha Institute).