barcodes on both ends of amplicons?

[ajthomas]

I'm working on designing a project to detect SNPs in a large number of samples. There are 6 SNPs (amplicons) to test in ~5000 samples. I have done a lot of amplicon sequencing on the 454, and with large numbers of samples I have used barcodes on both ends of the amplicon and used the combination of barcodes to identify samples, drastically reducing the number of primers needed. 454's AVA software has the ability to do that, but I've never used anything else to separate reads. In this case I don't need long sequences, so I want to use another sequencing platform to reduce the cost. A single run on a MiSeq or a Proton or a PGM would be sufficient as opposed to 2-3 runs on the 454.

So, here's my question: is it possible to use barcodes on both ends of the sequence using other sequencing technologies? For example, if I were to use paired-end reads on a MiSeq to get both ends or use an Ion Torrent system and keep the amplicons short enough to read clear through, are there tools out there that can look at both ends of the sequence for barcode information to separate the sequences?

[GenoMax]

http://www.illumina.com/products/nextera_xt_dna_sample_prep_kit.ilmn

http://www.illumina.com/Documents/products/datasheets/datasheet_nextera_xt_dna_sample_prep.pdf

MiSeq will automatically demultiplex the samples for you.

[ajthomas]

As far as I know, Illumina libraries typically only have the barcode on one end, and therefore I wouldn't expect the Illumina software to be capable of using the combination of two different barcodes, one on each end, to demultiplex. Are you saying that it can do that?

[GenoMax]

As far as I know, Illumina libraries typically only have the barcode on one end, and therefore I wouldn't expect the Illumina software to be capable of using the combination of two different barcodes, one on each end, to demultiplex. Are you saying that it can do that?

Simple answer .. yes. We have been running dual-indexed libraries for some time now.

Did you check the two links I included in the post above?

Illumina officially supports 96 indexes (8 x 12) but there are other threads on using more than that number.

[SNPsaurus]

ajthomas, you mentioned doing paired-end reads to get both ends of the fragment (and the barcodes), and sequencing clear through to get both barcodes, but it is a 454-style of barcoding to put the barcodes in the same read as the genomic information. Illumina now does separate reads for the indexes, and the barcodes reside in the adapter without needing to flank the genomic sequence.

When I developed RAD-Seq, we used "inline" barcodes that were part of the read, as the barcoding technology from Illumina wasn't suitable for large numbers of samples within a lane. But we've switched to the dual-index barcodes for nextRAD, as that supports much larger numbers of samples within a lane.

[RCJK]

Here's a protocol demonstrating how to use dual indexing with amplicons on MiSeq:

http://www.mothur.org/w/images/0/0c/Wet-lab_MiSeq_SOP.pdf

[JackieBadger]

MiSeq can sequence 2 x 250bp paired-ends. So if the total length of your target+primers+barcodes is < 500bp you can get the whole amplicon by sequencing in from either end and building a contig on the region of overlap.
No doubt you used the Roche MIDS before?
We do the same on the MiSeq, however I wouldn't leave any demultiplexing up to the MiSeq on machine software as it is pretty crappy. You could use Illumina's CASAVA which will de-multiplex, but easier still (and what I use) is jMHC.

Be aware that what people have referred to here as indices are not the same as MID barcodes (although can effectively be used in a similar fashion). If you already have your MID-labelled primers then you of course do not need do use duel-indecies and you can prep your amplicons the same as you did prior to 454 library prep. By adding in an Illumina index to your MIDs you get further multiplexing capabilities. the one issue you will have with amplicon sequencing on the MiSeq and many samples is the common problem of low-complexity. This means you have to spike you library with high complexity sequence.