[creeves]

Bar codes

The 8-base i5 and i7 indices had a Hamming distance of at least three from all others.

[koadman]

Well done creeves on pushing the multiplexing so far. What was the edit distance between your barcodes? The default parameters for demultiplexing tolerate 1 mismatch in each barcode I believe. If some of the barcodes have an edit distance less than 2 you might need to demultiplex looking for exact barcode matches.

[GenoMax]

However, we have since tried running over 4000 samples in a pool and here the MiSeq had big trouble demultiplexing. Has anyone out there tried such high levels of multiplexing? If so, I would like to hear about your experience.

You should be doing the demultiplexing off-line using bcl2fastq/CASAVA when you are doing something so outside "normal"

What was the length of the tag reads BTW.

[creeves]

High multiplexing of Nextera libraries

I would like to reopen this thread, as we have successfully run a pool of over 1400 samples (plasmids) on a MiSeq using our own i5 and i7 index primers. We reduced the volume of tagmentation considerably and developed a PCR protocol to attach unique combinations of our own i5 and i7 primers to each sample. The MiSeq had no trouble demultiplexing these 1400 samples. We hope to publish the methods soon. However, we have since tried running over 4000 samples in a pool and here the MiSeq had big trouble demultiplexing. Has anyone out there tried such high levels of multiplexing? If so, I would like to hear about your experience.

[TonyBrooks]

1) You should be able to design an i5 index primer/adaptor
2) The i5 index is primed and read from the P5' oligo that's grafted onto the flowcell for clustering. No custom read primer is needed. However, due to the way linearisation is performed during cluster-gen. 7 dark cycles are needed to extend this oligo until the i5 index is reached. See page 7 of this document (http://supportres.illumina.com/docum...15032071_b.pdf)
3) Read sequence goes as follows: Read 1, i7 index, i5 index, Read 2. Everything but read 2 is read off of one strand (to allow for dual indexing on single read runs without the need to resynthesise the second strand).

[krispy]

Hi all,

I am not sure if I should ask it here, still new to NGS, and I have lots of questions that I couldn't figured out.

1) I am using a custom adapter (that contains p5 p7 for flow cell binding), custom sequencing primer and also a custom index i7 primer for my work. I come across this dual-indexing, I am just wondering if it is possible to have a custom index i5 primer?

2) I am not familiar with Nextera, why is the dark cycles necessary? why not just custom make an index i5 primer that spam the 7bp constant read and read directly into the index? or wait did I miss out something. where is this index i5 primer primed from?

3) must it be index read 1 = i7 index; index read 2 = i5 index? can't be other way round? it's still on the same strand right?

sorry if I asked too many of the dumb questions.

cheers.

[pmiguel]

Okay, my mistake. Your adapters should be fine.
Is this just a normal NexteraXT kit library, or a set of amplicons you constructed in another manner to mimic the structure of a Nextera library?

We just did a MiSeq run on a NexteraXT library (our first Nextera run). It worked. I even spiked in 30% phiX, just in case. (Denature the normal, NaOH way, diluted to 8pM and mixed 30:70 vol:vol with the heat denatured NexteraXT library.)

Actually, that was our second Nextera run. We ran the same library (no phiX) the previous day on another MiSeq and it terminated prematurely (crashed) without imaging apparently. Maybe I should have mentioned this earlier? But I was pretty sure this was a fluidics issue because the instrument failed a subsequent volume test.

--
Phillip

[uberfinch]

Yes, I'm positive it is a Nextera XT kit (just bought it!). If you look on the illumina letter, those sequences are from the Illumina V.2 nextera kits.

Anyway I guess there isn't a good explanation for why it didn't work. Maybe I'll just have to risk burning another kit! There's only so many times you can stare at the sequences looking for an error.

[pmiguel]

To make absolutely sure it was right I TOPO cloned a library made with the Nextera XT kit that is supposedly compatible with the miseq on got that exact adapter sequence.

Are you sure it was a Nextera XT kit, not an old (possibly Epicentre) Nextera kit that you topo cloned and Sanger sequenced?

--
Phillip

[pmiguel]

PCR primers? No.
It is the internal segment, closest to the "insert" that has been replaced in the Illumina Nextera kits.
Details in the "Illumina Customer Sequence Letter".
Again, I think you probably did get clusters on your flow cell. They just were invisible because Illumina sequencers only detect clusters to which a read1 sequencing primers can anneal.

--
Phillip

[uberfinch]

Thanks for your reply Philip. Yes, as you surmised those 19 bp on the end are the nextera sequence taken from the sequence reported in Illumina's letter for version 2. To make absolutely sure it was right I TOPO cloned a library made with the Nextera XT kit that is supposedly compatible with the miseq on got that exact adapter sequence.

Is there any reason to believe the PCR primers for nextera are somehow modified?

[pmiguel]

I used the following adapters on our MiSeq:

N701:
CAAGCAGAAGACGGCATACGAGATTCGCCTTAGTCTCGTGGGCTCGGAGATGTGTATAAGAGACAG

N501:
AATGATACGGCGACCACCGAGATCTACACTAGATCGCTCGTCGGCAGCGTCAGATGTGTATAAGAGACAG

It didn't detect clusters (I used the provided sequencing primers).

Anyone have any ideas of what went wrong? I'm dumbfounded.

I guess I don't understand your strategy.
Looks to me like clusters should have formed, but I don't see anywhere for the standard sequencing primers to anneal. Which would make the clusters undetectable unless you added custom sequencing primer. If you did add custom sequencing primer, does it have an annealing temp of 65 oC (or higher)? The Illumina document that Tony attached above points out that primers with lower annealing temps might not work.

Eg, your i5 adapter:

AATGATACGGCGACCACCGAGATCTACA | CTAGATCG | CTCGTCGGCAGCGTCAGATGTGTATAAGAGACAG
Flow cell binding----------- | index??? | ?____________________what is this?


Oh wait, that is the old Nextera transposase sequence. Guess Illumina does not include it in their standard i5 primer cocktail. You would have to spike it in.

--
Phillip

[uberfinch]

MiSeq fail with Nextera adapters

I used the following adapters on our MiSeq:

N701:
CAAGCAGAAGACGGCATACGAGATTCGCCTTAGTCTCGTGGGCTCGGAGATGTGTATAAGAGACAG

N501:
AATGATACGGCGACCACCGAGATCTACACTAGATCGCTCGTCGGCAGCGTCAGATGTGTATAAGAGACAG

It didn't detect clusters (I used the provided sequencing primers).

Anyone have any ideas of what went wrong? I'm dumbfounded.

[TonyBrooks]

See this document (taken from an Illumina bulletin on their website).

[gnomers]

I think both should work.
The Illumina read primer is actually a primer mix of several primers. I guess there's probably primers for both sequences in there (TruSeq and Nextera).

However, if you're generating library directly from PCR, why not remove those sequences all together and just use custom read primers. Your custom read primers will just be versions of your fwd and rev primer sequences. You'll also waste less reagent by not sequencing your primers. See my post further up.

We've been using custom read primers on an amplicon library successfully on the HiSeq. However, we just tried the exact same strategy on the MiSeq and it failed utterly (nothing but PhiX in the first (index) read.)

Are the annealing temperatures very different on the HiSeq and MiSeq?