Demultiplexing dual-indexed MiSeq fastq files

[lynchde]

Hi all. I received .fastq files from an overlapping-paires paired-end MiSeq run. I joined the pairs together using FLASH. However we used dual-indexing, so now I've got sequences that look like this

5' barcode1-link-primer1-----sequence-----primer2-link-barcode2 3'

(actually its more like
5' barcode1-link-primer1-----sequence-----2remirp-link-2edocrab 3' because of the reverse & join!)

So I need to demultiplex these (including primer removal). Does anybody know of any good programs that can demultiplex dual-indexed .fastq files? Any and all help much appreciated.

[lorendarith]

Sorry, but something is not right here or I'm not quite following...

If you did Illumina dual indexing you should have the reads and indexes in 4 separate (FASTQ) files (read 1, 2, 3 and 4). You should not have any indexes in your reads, unless you did additional barcoding?

If you did, then the barcode is supposed to be after the primer site (where the sequencing primer binds) or else it would not get sequenced.

Here's a handy pic of standard dual indexing:


In case you did barcoding, you only want to remove the barcode from the beginning of each read, which you can do with almost any trimmer, before overlapping the reads. If you already overlapped you can use something like http://chipster.csc.fi/manual/prinseq-trimmer.html

Here's a demultiplexer for dual indexed reads:

https://github.com/sebhtml/FastDemultiplexer

[lynchde]

Thanks lorendarith.

Sorry. That's my fault. I used the word "primer" instead of "adaptor". So it should really read:

barcode1-link-adaptor1-----sequence-----adaptor2-link-barcode2 3'

We used our own barcodes, not Illumina.

All I've got is 2 files, for reads 1 and 2 (no 3 or 4).

The problem is that both barcodes are required to identify which sample a read came from (i.e. beginning for readfile1 AND beginning of readfile2). So if I trim before merging the two reads, I need to make sure that I keep all of the pairing information so that I can still tell which samples they are from. So I'm not sure these programs will help but I'll take a closer look.
Thank you.