Advice on analyzing barcoded samples with GERALD/ELAND

[lparsons]

Does anyone have any advice on the bioinformatics side of dealing with barcoded data? Specifically, we are using the Illumina Pipeline 1.3.2.

I would like to be able to separate the reads based on barcodes and then align them using GERALD/ELAND. It seems I could use just the ELAND_standalone.pl script, but that would not give me the added "benefits" of GERALD such as quality value recalibration and more improtantly, the "extended" part of eland_extended to extend alignments past the first 32 bp.

More generally, I'd like to be able to use ELAND to align reads from FASTQ files from various sources, but that doesn't appear very feasible at the moment.

[cbrennan]

If you know the length of your barcodes -- use the USE_BASES option the strip the barcode out of the Eland alignment.
For ex: if your barcode is 5 bases: USE_BASES nnnnnY*n

It will ignore the first 5 bases & then align. When the analysis is done you should be able to parse the eland output files & bin them by the barcode.

**When you bin, be sure to allow for indexing errors - you may have lost the first base of your barcode.

If you're using the Illumina barcode kit -- when you upgrade to v1.4 you should be able to use the 'indexing' option in the USE_BASES directive and the pipeline should separate the barcodes out for you -- haven't tried this myself yet.

[elaney_k]

It will ignore the first 5 bases & then align. When the analysis is done you should be able to parse the eland output files & bin them by the barcode.

Hi cbrennan,

Which files in the eland output contain the full-length read (i.e. index + read) when you use the USE_BASES nnnnnnY*n option?

I index my samples but when i run eland with the USE_BASES nnnnnnY*n option all of the output files contain reads minus the first 6 bases. At the minute I run eland twice; once with USE_BASES all and ANALYSIS sequence to get the full length reads that i can subsequently parse by index and the other with USE_BASES nnnnnnY*n and ANALYSIS eland_extended to get the summary statistics for the run.

Elaine

[elaney_k]

Actually I've just figured it out, instead of just using the USE_BASES nnnnnY*n option you need to use the I6Y*n option where "I" indicates an indexed base and 6 being the number of bases in the index.
The eland output will then have the index in the header for each read or in the export file it will be listed in column 7.