demultiplexing using fastq-multx

[Mark Roberts]

Hi,

I am using the following command for processing illumina MiSeq data:

fastq-multx -g I1.fastq R1.fastq R2.fastq -o r1%.fastq r2%.fastq

I1.fastq is the illumina index read, R1 and R2 are obviously the paired-end reads.

The command is returning this error:

File I1.fastq isn't a barcode-only file, try using -l instead

Not sure what to do about this?  R1 and R2 don't have the barcode in the header or the body of the sequence.  I imagine using the index to determine which barcode to assign which sequence to based on matching headers.  If this script matches barcode sequence in I1 to barcode in headers to demultiplex than I am out of luck.

Anyone have any ideas about either what I am doing wrong or what tool would be appropriate.

Please excuse my ignorance, this is my first nextgen run.

Thanks!

[Erik Aronesty]

fastq-multx runs in several modes.   

the simplest mode is with "-B BCFIL"... where you list the barcodes in a separate file, and then give all 3 reads from illumina.   this would be the way than most people run it, and is the easiest to diagnose.

with -g... the first file "l1.fastq" is assumed to be the index read.... a short read less or equal to than 11 bp per read.   statistics are then used to auto-determine the most prevalent (10% of max or greater) sequences in that file, and then use only those when demultiplexing.   this is a great method when the input is ill-specified, or in a "general-purpose" pipeline that remains adaptable to changing inputs.

in this case, the error is that, specifically, the file you provided has an average length >= 12.... is it really the index read?   if so, what's the average length?