Single or Dual indices?

[farrel75]

Hi all,
I am new to Illumina and am preparing a library of 24 samples with the Nextera XT. I plan on runing these samples on the MiSeq. I've read a couple threads where people mention single or dual indices such as http://seqanswers.com/forums/showthread.php?t=14676.
Why you would do one over the other? Does it make a difference?
Thanks!

[kcchan]

Nextera XT uses a dual index strategy which allows 24 unique barcodes using only 10 oligos (4 of Index 1 and 6 of Index 2) instead of 24 unique oligos for each barcode. You can mix and match dual and single indexed libraries in a sequencing run, but not for the sample prep. For Nextera you must use both index oligos for your sample preparation.

[pmiguel]

Just to be clear, if you use a Nextera kit, you get dual-indexed amplicons. You don't have to sequence both indexes (or any of them), but they are there, none the less.

Since you are doing a MiSeq run, I would suggest running dual indexing since that is built into the MiSeq reagents, anyway. But for 24 indexes, you would have to deploy duals anyway, because the standard Nextera kit has a maximum of 12 single indexes. Of course you could "go rogue" and synthesize your own step-out primers to get you up to 24 single indexes. But the cheap Nextera adapter kit offers 24 index pairs, so why look further?

--
Phillip

[farrel75]

Thanks everyone! It's good to know that its recommended to use dual indexes for the MiSeq and Nextera XT. I guess I didn't make it clear that I was wondering why in the general sense would you do single vs. dual indexing. What are the benefits/problems of one index vs. two?

[pmiguel]

Dual costs you an extra 8 cycles + some dark cycles. But MiSeq reagent kits are rated for how many sequence (insert) sequence bases you read, with enough extra so that you can do all the extra cycles required for dual indexing.

So, the only reasons I can think of to choose single index, would be if you really wanted your reads to be a few bases longer or you could not afford the extra hour or two required to run the second index.

--
Phillip

[austinso]

Off-hand does anyone know if you can specify reading only the second index "i5" and not the first "i7"?

[kcchan]

Off-hand does anyone know if you can specify reading only the second index "i5" and not the first "i7"?

It might take some hacking of the recipes to get that to work. A simpler workaround would be to use the P5 primer as the index 1 primer and then skip the "index2" read. You will have to burn an extra 7 cycles to get through the padding sequences before you start reading the barcodes, however.

[austinso]

Thanks for confirming that, kcchan.

[HeinKey]

Hi Austinso,
If you use low complexity pooling strategies be aware to have sufficient balance in each index cycle.
Choose the indices properly, making sure the combination has 50% A/C and 50% G/T per cycle. Strong bias in variation is easy with only a few different indices. Basecalling can be very much off with biased indices.
The latest MCS version has a new algorithm and works much better with biased libraries though.
It's a miracle to me they don't implement this algorithm into the next HCS update also. "No that's not planned" was Illumina's reaction when I informt for this.