Demultiplexing PE Illumina reads

[Girma]

Hi there,

Recently, we made PE Illumina sequencing of multiplexed libraries (~320 bp and up to 23 libraries per lane). Target libraries were prepared with single 6-bp Illumina index (not dual), and the cluster was generated using PE cluster kit v3.

Unfortunately, during the sequencing, the technician did not specified about indexing. Now I generate sample sheet (describing the samples and indexes) and tried to demultiplex and regenerate Fastq files with Cassava 1.8.2. But the majority of the reads were in ‘unspecified indices’ directory. I tried to adjust –use-base-mask option, but no changes. The other problem I had with –use-base-mask setting is that the length of each read is 104 bp (R1 & R1 = 204, which is also specified in both Config & RunInfo files), and when I set –use-base-mask as Y*n,I6n,Y*n or Y*,I6,Y*, it set the index containing reads to be R2, but still no changes. I am not sure if my –use-base-mask setting is right. When I run Cassava with default settings, error message saying ‘expected barcode length –including delimiters is 0’ and the program terminated. Please any helps and suggestions on how to fix this?

I also tried other three tools – Fastx, EA-utils and Sabre – but they are not good enough (majority of the reads without index).

I enclosed the top-reads generated from the original settings (without indexing).

Thanks!

[GenoMax]

If this run was set up by the tech as a plain PE run (i.e. no multiplexing) then is nothing you can do but re-run. In illumina technology tags are read as independent reads and if they are absent (if example you posted is all you see in the undetermined reads pool) then one can't demultiplex the data.

If you had requested that the sample be run as multiplexed then it is the fault of the facility and they should re-run your sample for free. If you did not specify in your submission that the sample was multiplexed then you would need to absorb the cost of the re-run.

[Girma]

Thank you GenoMax!

You are right and I also discussed with Illumina people that I cannot demultiplex my data – since index read was not performed . It’s very unfortunate and it’s not my fault.

[GenoMax]

It is frustrating when something like this happens. Perfectly good sequence data (in terms of quality) but of no use.