[Vinz]

We are using the 500 cycle kits for amplicon sequencing with about 20 different amplicons. After applying hardcoded settings it looks ok now. Not as nice as a phiX. But we get usually above 75%Q30. The last 50 bases of the reverse read quality drops (20 to 50% below Q20).

[LVAndrews]

Also, it was my understanding that the new MCS version 2.1.1.13 had some improvements for handling low diversity data. Anyone seen evidence of this?

--
Phillip

I had thought this as well, but my first amplicon run after the update didn't seem too great. I asked our FAS and she said it was mainly meant to prevent the intensity spikes that result from high phasing/prephasing values, not to really address the low-complexity problem.

[LVAndrews]

We are using the 500 cycle kits for amplicon sequencing with about 20 different amplicons. After applying hardcoded settings it looks ok now. Not as nice as a phiX. But we get usually above 75%Q30. The last 50 bases of the reverse read quality drops (20 to 50% below Q20).

Vinz, are you using barcoded primers for sample prep or do you have universal adapters instead? I am interested in using the Y-adapters from the Kozarewa and Turner protocol for T/A cloning amplicons since this would make our adapters universal for any amplicon, but I'm not sure if template orientation will be a problem in downstream analysis since at least with barcoded primers, orientation is always the same whereas with Kozarewa adapters, the reads will go either way. This would have the added benefit of increasing complexity and one could even move the index to the front of read 1 for 12 cycles to increase complexity during the initial part of the run. The Caporaso paper insisted on a separate index read mainly because Hiseq reads were so short, but in the era of 300 and 500 cycle kits this seems less important to me now.

[Vinz]

We use the 4 primer approach with dual indexing (see this post). Meaning, we do only 1 PCR which includes a target specific PCR including universal adapters and outer primers which will add the indexes and flow cell sequences. I guess it depends on how many different amplicons you want to target. If you have a low level of targets and multiplexing, then your approach might be worth while.

[LVAndrews]

We use the 4 primer approach with dual indexing (see this post). Meaning, we do only 1 PCR which includes a target specific PCR including universal adapters and outer primers which will add the indexes and flow cell sequences. I guess it depends on how many different amplicons you want to target. If you have a low level of targets and multiplexing, then your approach might be worth while.

That's fantastic. Exactly what I've been looking for. We're new to providing miseq service and it seems like everyone wants a different locus (which I totally understand).

[boscoliveira]

That's fantastic. Exactly what I've been looking for. We're new to providing miseq service and it seems like everyone wants a different locus (which I totally understand).

Hi Vinz, would the following primers work (including 12N to try to solve the low diversity problem), Nextera style (using the transposase sequence available in the Illumina Letter):

PCR1F TCGTCGGCAGCGTCAGATGTGTATAAGAGACAGNNNNNNNNNNNNSEQUENCESPECIFICPRIMER
PCR1R GTCTCGTGGGCTCGGAGATGTGTATAAGAGACAGNNNNNNNNNNNNSEQUENCESPECIFICPRIMER

PCR2F AATGATACGGCGACCACCGAGATCTACAC[i5]TCGTCGGCAGCGTCAGATGTGTATAAGAGACAG

PCR2R CAAGCAGAAGACGGCATACGAGAT[i7]GTCTCGTGGGCTCGGAGATGTGTATAAGAGACAG

[TonyBrooks]

I had thought this as well, but my first amplicon run after the update didn't seem too great. I asked our FAS and she said it was mainly meant to prevent the intensity spikes that result from high phasing/prephasing values, not to really address the low-complexity problem.

Apparently the new version of RTA is significantly better at low diversity. Phasing calculations are done empirically for every cycle, whereas previously the first 12 cycles were used as an estimate for the whole run.
Colour correction is now also estimated over more cycles too making the matrix a bit more robust.
Our area rep presented some data on this recently with a 900k/mm2 loading density, 90% PF with <5% PhiX. It comes with the warning that this isn't a magic fix. It still depends a little bit on what's exactly in your library but it does seem to improve data quality for single amplicon libraries.

We're yet to give it a go. We're thinking of starting with a 15% PhiX spike-in and titrating down as we go.

[Vinz]

I have not tried nextera style sequencing primers. All I can say is that until the indexes the sequence looks good.
Our attempts with 12 Ns were not too encouraging. We had the impression that introducing the 12 Ns resulted in more artefacts (short dimers).
Anyway, as Tony stated, the new software works really well. We spike in around 10% phiX with good success. Hardcoded phasing/matrix does not seem to be required anymore. So I would not recommend the 12 N approach.

[GW_OK]

You guys should really try phase-shifting-N primers. Have a few sets of primers with varying numbers of N before your amplicon (N, NN, NNN). This gets the identical regions of each sample out of sequence phase with each other. Works well for us so far.