Demultiplexing in-line barcodes

[saikumarkv]

Hi all,

I just used a kit that has in-line barcodes for multiplexing RNA-seq samples. CASAVA cannot handle in-line barcodes. Can anyone suggest if there are any other tools and the best way to demultiplex in-line barcoded data.

Thanks

[HESmith]

CASAVA can handle in-line barcodes just fine. Check out the user's guide.

[saikumarkv]

HESmith, when I talked to Illumina they said it cannot handle in-line barcodes. I will check the manual. Thanks
Sai

[kmcarr]

HESmith, when I talked to Illumina they said it cannot handle in-line barcodes. I will check the manual. Thanks
Sai

They lied. Or to be more polite, the representative who spoke to you perhaps does not fully understand the capabilities of the CASAVA software which led him/her to answer erroneously.

I have demultiplexed in-line barcodes using the latest version of CASAVA. You just need to properly configure the "use-bases-mask" parameter when you run configureBclToFastq.pl. For this parameter the different reads are separated by commas and the use of bases by 'y', 'n' or 'i'. y==this base is part of the read, n==don't include this base in the output and i==this base is part of the index.

You didn't specify how you constructed your library so I will give you an example based on our experience. We designed custom "Y" adapters which placed a 3bp barcode at the ends of the fragment; this would be immediately followed by a 'T' (due to 'A' tail ligation) which we would want to skip and then we are in to the read. We sequenced these fragments with a 2x101bp run. The library construction method would mean that the identical barcode should be present at the 5' end of both reads but we only used one (at the time CASAVA only allowed 1 barcode). To configure CASAVA I used:

--use-bases-mask i3ny*n,n4y*n

This string tells CASAVA:

• Use the first three bases of read 1 as the index (configured in SampleSheet.csv).
• Skip the next base.
• Use the remaining bases (except the last one) as sequence read 1.
• Skip the first 4 bases of read 2.
• Use the remaining bases (except the last one) as sequence read 2.

[gruberjd]

Sai-- there are a number of post-casava tools (mostly free) that can do this. For example, the fastx toolkit's barcode splitter works:

http://hannonlab.cshl.edu/fastx_toolkit/commandline.html#fastx_barcode_splitter_usage

(apparently you need to click this, it won't display the whole address)

A complication that is not solved by the fastx toolkit is paired-ends, but if that isn't relevant to you, it will work well.

[saikumarkv]

Demultiplexing in-line barcodes

You didn't specify how you constructed your library so I will give you an example based on our experience. We designed custom "Y" adapters which placed a 3bp barcode at the ends of the fragment; this would be immediately followed by a 'T' (due to 'A' tail ligation) which we would want to skip and then we are in to the read. We sequenced these fragments with a 2x101bp run. The library construction method would mean that the identical barcode should be present at the 5' end of both reads but we only used one (at the time CASAVA only allowed 1 barcode). To configure CASAVA I used:

kmcarr: I used NuGen's Ovation pico kit (older version) that has 4b barcodes. Thanks for the info. I will pass it on to our informatics people. They typically get data from TruSeq library samples. This is the first time they are analyzing data with inline barcodes. I really appreciate your input.

[saikumarkv]

Demultiplexing in-line barcodes

Sai-- there are a number of post-casava tools (mostly free) that can do this. For example, the fastx toolkit's barcode splitter works:

http://hannonlab.cshl.edu/fastx_toolkit/commandline.html#fastx_barcode_splitter_usage

(apparently you need to click this, it won't display the whole address)

A complication that is not solved by the fastx toolkit is paired-ends, but if that isn't relevant to you, it will work well.

Gruberjd: Thanks for the info. I run the sequencing core at University of Cincinnati. Only recently I'm getting into data analysis.