Split barcodes - embedded index only in forward read

[memento]

Hi,
I try to split fastq file containing multiple samples (illumina paired end, miseq run). These are not general barcodes read directly on the machine - they are simply embedded within the read (forward only, first 6bp, reverse has no index).
Tried to use FASTX toolkit but there is no way to use it with paired end.
Managed to successfuly split the read1 using FASTX but how to match read2 reads ?
Is there any tool to directly work on PE reads?
Thanks in advance!
b

[JackieBadger]

Your problem is one I image a lot of people have yet there isn't a straight forward way of doing what you ask (that I am aware of).

One way to do it is to filter you R1 and then interlace each filtered file with its matching R2. For this I use FastqInterlacer/de-interlacer implemented in Galaxy. You will need to install a local instance of Galaxy and then add this functionality via their toolshed.

Another way that may work is to use the RADseq program STACKS. The QC step in STACKS (process_radtags) splits paired end reads by barcodes in R1. Just make sure you switch off all off the other QC filtered to keep all of your data.

[nsaarman]

https://github.com/najoshi/sabre