Custom dual index adapters...dark cycles unnecessary?

[koadman]

I am wondering about dual index adapter designs that are compatible with the MiSeq's standard dual index sequencing protocol on paired-end flowcells. The protocol seems to roughly be (apologies for sloppy terminology):

• read 1
• index 1
• bridge
• 7 dark cycles
• index 2
• reverse strand synthesis
• read 2

Does anybody know or have ideas about whether the 7 dark cycles for sequencing a dual-index library on a paired-end flowcell can be eliminated by using custom adapter sequences without those 7 bases? What is the rationale for an adapter design that needs the 7 dark cycles? Is it only so that libraries can be sequenced on both single- and paired-end flowcells?

I would like to reduce the total cycle count for dual indexing to be within range of the standard HiSeq 2000 kits and avoid the need to top up the reagent cartridges.

[ECO]

Wondering if you tried this. If not I'll be trying it or a variant of it soon.

[koadman]

I didn't try it yet, we're just using the instrument's standard dual-index program and adapter design. If you try it I would love to hear about the experience. I will say though, that my first attempt at working out the position for the 2nd index in the adapter wasn't quite right. Fortunately we only needed the first index to demultiplex that particular pool of samples.

[steinmann]

Wanted to design custom dual index adapters for Nextera libraries but Illumina tech support would not give us the sequences of the adapters. From this discussion I assume that you guys were more lucky?

[koadman]

Nope, we didn't get those from Illumina either, and not for lack of trying. It seems like someone who has done a nextera run with the new kits might have enough adapter read-through to get these unless they did really stringent size selection. If anybody thinks they may have such a run they're willing to share I would like to have a go at extracting those sequences.

My approach has been to modify the Nextera v1 adapters to add the 2nd index. In a stroke of brilliance Epicentre included the adapter design in the protocol that came with its kits so that users could synthesize arbitrary barcodes. Seems unfortunate that Illumina has not been so open with the v2 kits.

[steinmann]

Thats really unfortunate. Tried the sequences from the Epicenter kit but PCR failed. Did you get amplification of you library with your self-syntesized adapters? Have now ordered an Illumina adapter kit and will clone my library so I can sanger sequence a few fragments for QC and to see what the adapter sequences are.

[koadman]

Yes, we get library amplification when using 3rd party synthesized adapters with the Epicentre Nextera v1 kits. We have seen the nextera kit fail to amplify in some cases. This sounds vague, but some samples seem to contain something that inhibits transposase activity. Wish I could be more specific. Other culprits could be highly degraded starting material (low molecular weight), or not enough material, or too much material. Flourimetry-based quantification is essential, methods like the nanodrop seem to grossly overestimate the amount of DNA.

[steinmann]

Thanks for the info! My experiences were with the Illumina v2 kit. Dont thinkt ammount of sample was the issue. Guess I will see when my Illumina adapters finally arrive. Will let you know how that went if you are interested.

[steinmann]

Does anybody know or have ideas about whether the 7 dark cycles for sequencing a dual-index library on a paired-end flowcell can be eliminated by using custom adapter sequences without those 7 bases? What is the rationale for an adapter design that needs the 7 dark cycles? Is it only so that libraries can be sequenced on both single- and paired-end flowcells?

If you leave out those 7 bp primer 2 wont be able to bind and you can not amplify your library. So you would also have to shorten primer 2 by the 4 bp that anneal in this 7 bp region.

No clue what complications besides a 5 degree celsius drop in the melting temperature of the primer this could lead to.

[Vinz]

If you leave out those 7 bp primer 2 wont be able to bind and you can not amplify your library. So you would also have to shorten primer 2 by the 4 bp that anneal in this 7 bp region.

No clue what complications besides a 5 degree celsius drop in the melting temperature of the primer this could lead to.

The 7 bases are the bases of the flow cell adapter that are cleaved off by an restriction enzyme after bridge PCR (before read1). If you leave those out the end of the adapter sequence is not complementary to your product. That will lead to very bad bridge PCR efficiency. So unless you design your own flowcell or modify the flowcell from illumina, there is no way to safe on these 7 cycles.
(The primer for sequencing of read 1 and read 2 are not affected by changes of the sequence outside the barcodes.)

[koadman]

Hmm, that would seem to kill the idea of saving cycles. Will be interesting to hear whether ECO has some idea to circumvent this?

[steinmann]

What has been done is to cleave off the nextera sequences using restriction enzymes and then ligate adapters with inline barcodes.

http://genome.cshlp.org/content/early/2011/11/16/gr.124016.111

For me thats not an option but maybe it makes sense for you.

[Igor]

It seems to me that it should be possible to use the strategy employed for reading the second index (i5) on single read flowcells. That is, perform read one and i7 index reads, then use index 2 sequencing primer (HP9) provided with the dual index sequencing primer kit for SR flowcells to read index i5, then perform second strand resynthesis and linearization as done during the regular (single index) paired end protocol, and finish up with the second read. No dark cycles needed. I am probably missing something obvious...

[Vinz]

It seems to me that it should be possible to use the strategy employed for reading the second index (i5) on single read flowcells. That is, perform read one and i7 index reads, then use index 2 sequencing primer (HP9) provided with the dual index sequencing primer kit for SR flowcells to read index i5, then perform second strand resynthesis and linearization as done during the regular (single index) paired end protocol, and finish up with the second read. No dark cycles needed. I am probably missing something obvious...

I see no reason that should not work. Good idea!