Critical operational note for those using this protocol: As currently written, this protocol is not appropriate for new Illumina sequencing technology (i.e. patterned flow cells found in NovaSeq). This is due to more serious index-hopping artifacts that are found in this new version of their platform. We will upload an updated protocol, with unique dual indexing in the primers, as soon as we have fully tested it.
As currently written, this protocol is still appropriate with older platforms such as HiSeq 2500, and we recommend Tufts’ Paired-End 250 Rapid-Run Service (RL-PE250) service for our US collaborators (Website with price info). However, this platform is being phased out so may not be available for much longer. Moving forward, MiSeq will be the only platform to support 2x250bp reads that could support the currently version of the protocol.
We use a “DIY” protocol for preparing amplicons that produces high-quality, quantitative data. We use long primers with Illumina adapters already as part of the primer, meaning that there is no additional adapter ligation step. This ligation step might have some negative effect on the quantitative nature of the PCR assay but it also complicates sample preparation. The only downside of our way of doing things is you have to be really careful to note which forward and reverse primers correspond to which sample. After your PCR is completed, all you need to do is to cleanup with Ampure XP beads, quantify, and pool at equimolar concentrations.
The latest and greatest version of the protocol can be found here:
https://www.protocols.io/view/fuhrman-lab-515f-926r-16s-and-18s-rrna-gene-sequen-vb7e2rn
Important Notes on Calculating Sequencing Depth
Thank you to Liv for sharing her experience in this matter!
We normally sequence our cleaned amplicons on a 2x250 HiSeq RapidRun which gives us ~180 million reads. To provide the requisite “random” DNA that the sequencer needs, we pool our amplicons with metagenomes. The calculation for this pooling is somewhat counterintuitive, but this is how it has worked in the past:
- Create pool that contains 75% amplicons by molarity; the remainder is metagenomes.
- Put on sequencer, which yields 180 million reads on average.
- For whatever reason, there is an approximate 3x bias against the amplicons, which means you get approximately 25% of the total 180 million reads back as amplicons, and 75% is the metagenome you added (why? Bob only knows…).
- If you add approximately 280 amplicon samples, you will get on the order of 100,000 reads/sample (though there is considerable stochasticity in this number).
- For our large size fraction (1.2 - 80uM), you can expect to get approximately 10,000 18S reads, 17,000 Chloroplast 16S and 73,000 other 16S (mostly Bacteria/Archaea but also including Mitochondria).
