FastQ files: Double demultiplexing

Question

[compuTE]

Hello,

Is there a way of demultiplexing Illumina Truseq RNAseq data more than once? For example, demultiplex by sample and another index? (this index being dual-index)

Of course, it's easy to demultiplex the samples using bcl2fastq. But in this case, I end up with a fastq file per sample containing reads with dual indexes that I want to split into different fastq files.

This is pared-end data.

I tried fastq-multx on the fastq files output by bcl2fastq, but most of my reads failed to be assigned to an index. It could very well be the case, but I would like to verify this approach somehow. I would appreciate any software recommendations.

Thank you!

[GenoMax]

You will need to explain where this second index is. Is is a standard Illumina second index? Probably not since in this case bcl2fastq would have handled it. Is it inside (in-line) your reads? If latter where is that located?

You may want to try sabre (LINK).

[compuTE]

Hi, thank you so much for your suggestion! The indexes are added at the 3' and 5' of the sequences and are complementary. For example, an index is placed as

ACAACC[-----an RNA molecule here-----]GGTTGT

So I think ACAACC should be at the start in either R1 or R2 and GGTTGT in the other one.

I tried sabre as you suggested and it seemed to have worked - I got 98% of the reads matched to one of my indexes. I used the following command:

sabre pe -m1 -c -f S1_R1.fastq.gz -r S1_R2.fastq.gz -b index_sheet_single.tab -u S1_R1_unknown.fastq.gz -w S1_R2_unknown.fastq.gz

And for index_shee_single.tab I have something like this (from the example):

ACAACC  group_1_R1.fastq    group_1_R2.fastq

I think (and please correct me if I'm wrong because i didn't fully understand how this worked..) -c from sabre looks for the reverse complement of the index in the other read mate.

[GenoMax]

If 98% of your data was demultiplexed then the tool did work ok.