I am working on a custom library preparation method and I designed my p5 oligos incorrectly. To be specific, I used the reverse complement of the correct p5 index sequence. As a result my fastq files aren't demultiplexing properly. Ideally, this should be an easy problem to fix, just switch the sequences in the sample sheet to match what I actually made, but for some reason DRAGEN's bcl conversion software refuses to run it and says there's an error with the sample sheet. I'm assuming it's because the index sequences don't match what it's supposed to be, but I'm well aware of that discrepancy and would like it if the sheet could just run. Illumina has been less than helpful on this front. Has anyone else ever had this issue and found a work around?
Forget DRAGEN in that case. Run the demux via
bcl-convertoffline. I assume you have access to full data folder.I do have access to the full data folder, but I would have to install bcl-convert on my group's local cluster and I've had a lot of trouble doing that previously because IT at my company is ridiculous and managing permissions is a giant hassle.
Get the undetermined read files (by using a blank samplesheet) and then demultiplex externally using
demuxbyname.shfrom BBTools orDeML(https://github.com/grenaud/deML ) ?We unfortunately do not have those programs either. We have BBMaps, but not BBTools.
BBMap is part of BBTools.
Ok, so it turns out when I was editing the sample sheet in excel (which of course was the problem) excel changed the NoLaneSplitting parameter from "true" to "TRUE" automatically. So there's the issue and now it's demultiplexing just fine. Never edit your sample sheets in excel folks.