Demultiplexing combinatorial barcodes from paired end data

[ddelalca]

Hey, I've been trying to figure out how to demultiplex paired-end 250 miseq data that has three different barcodes.

The samples all come from 96 well plates. I have two 5' barcodes representing plate number as well as plate column. These two barcodes are linked via a common adapter sequence. I then have one 3' barcode that represents the plate column.

I am fairly sure I would be able to demultiplexing the data in 2 rounds in which I first demultiplex based upon each plate/column combination, and then demultiplex the result based upon the 3' barcode.

This seems fairly cumbersome as I have 2880 unique samples. Is there some sort of software in which I can provide simply the three unique barcodes, and then demultiplex in one step?

Also, I usually use the A5Pipeline (https://code.google.com/p/ngopt/wiki/A5PipelineREADME) to error correct sequence reads, remove Illumina adapter sequences, and assemble the paired-reads. Would I be able to assemble the reads using A5 and then demultiplex the data, or do I have to run the pipeline on each individual sample after first demultiplexing. Thanks!

[ddelalca]

Does anyone have experience with combinatorial barcodes?

[GenoMax]

How long are these barcodes? It may be possible to write some custom code to do this in one step but without seeing what you reads look like (or a good schematic) no one is going to be able to help. It may be best to find someone locally who can work with you interactively.

If the A5Pipeline changes the context of the barcodes it will make things complicated. You may have to run the pipeline post demultiplexing.

[JackieBadger]

This will do it https://code.google.com/p/jmhc/

[GenoMax]

This will do it https://code.google.com/p/jmhc/

Can you import fastq files? Description states that it operates on fasta files (internally?). I don't think this will do 3 barcodes in a single pass.

[relipmoc]

use skewer

Hi ddelalca,
You may use skewer for this purpose. An example command line is as follows:

skewer -x V3-forward.txt -y V4-reverse.txt --mode ap -t 4 ampliconReads-pair1.fastq ampliconReads-pair2.fastq assigned/

If you only use some of the combinations, you may specify the matrix by -M as well.
Good luck!