Getting a list of all index sequences

[Mouth_Breather]

Hi!

For an internal project we are doing, we are trying to get at the actual index sequence for each read (all reads, whether it winds up in the undetermined indices bin or not).

We are using casava 1.8.2. From what I can tell looking at the Casava User Guide, this information is only present in the binary .bcl files.

Does anyone know of another way to retrieve this information, other than writing a script that parses binary?

Thanks for reading!

[kmcarr]

Mouth,

The actual barcode read for each read is recorded in the definition line of the FASTQ file. Here is the format of the Illumina FASTQ produced by CASAVA 1.8.x

@HWI-ST957:100:D0V52ACXX:6:1101:1221:2161 1:N:0:CGATGT

The barcode sequence at the end of the read is the actual barcode read for that cluster.

Here is an example showing that you can see variation in the barcode recorded in the defline. It grabs the first 1000 deflines from a gzipped fastq file, splits the defline at ":" and takes the 10th field (the barcode), sorts them and counts the number of each uniq one.

zgrep ^@HWI CTRL1_CGATGT_L006_R1_001.fastq.gz | head -1000 | cut -d":" -f10 | sort | uniq -c
      1 CGATAT
      2 CGATGA
      1 CGATGG
    987 CGATGT
      2 CGCTGT
      1 CGGTGT
      6 TGATGT

You can see that 98.7% match he expected and there are a few with mismatches.

But be aware if CASAVA demultiplexing was run with default settings no mismatches are allowed in the barcode. You will only see differences between the barcode read vs. the configured if you set up the CASAVA run (configureBclToFastq.pl) with --mismatches=1.

[Mouth_Breather]

Mouth,

The actual barcode read for each read is recorded in the definition line of the FASTQ file. Here is the format of the Illumina FASTQ produced by CASAVA 1.8.x

@HWI-ST957:100:D0V52ACXX:6:1101:1221:2161 1:N:0:CGATGT

The barcode sequence at the end of the read is the actual barcode read for that cluster.

Here is an example showing that you can see variation in the barcode recorded in the defline. It grabs the first 1000 deflines from a gzipped fastq file, splits the defline at ":" and takes the 10th field (the barcode), sorts them and counts the number of each uniq one.

zgrep ^@HWI CTRL1_CGATGT_L006_R1_001.fastq.gz | head -1000 | cut -d":" -f10 | sort | uniq -c
      1 CGATAT
      2 CGATGA
      1 CGATGG
    987 CGATGT
      2 CGCTGT
      1 CGGTGT
      6 TGATGT

You can see that 98.7% match he expected and there are a few with mismatches.

But be aware if CASAVA demultiplexing was run with default settings no mismatches are allowed in the barcode. You will only see differences between the barcode read vs. the configured if you set up the CASAVA run (configureBclToFastq.pl) with --mismatches=1.

Hi thanks for the reply! I'm aware of the ability to grab the barcode sequence from the fastq files - I was not aware that it showed the actual variations for those barcodes that have 1 mismatch - thanks for that.

But I also want to see the barcodes for which mismatches are 2 and greater. Are those recorded somewhere other than the .bcl files?

[kmcarr]

But I also want to see the barcodes for which mismatches are 2 and greater. Are those recorded somewhere other than the .bcl files?

Those are the reads in the fastq files under the Undetermined/Sample_lanex directories.