Support for UMI FASTQ

[adamrtalbot]

Description of feature

UMI data is sometimes stored in a third FASTQ file, typically because the UMI is embedded in the index and bcl2fastq2 cannot combine it into forward/reverse. This produces three files:
R1: Forward
R2: UMI
R3: Reverse

We can support UMI processing using this method. Key points:

• Allow the samplesheet to include optional 3rd FASTQ
• Ignore/handle UMI FASTQ when sending to modules that expect 2 FASTQ files to ensure compatibility
• Allow/enforce --umi_read_structure to support >2 masks.

See nf-core/fastquorum#11 for an example implementation. Could be implemented as part of NF-Core subworkflow.

[FriederikeHanssen]

Nice, yes I think this is already in the planning as soon as we have the shared subworkflow. ( I think there is a PR in modules repo somewhere, we should cross check it includes this)

[lauren-tjoeka]

Hi! Has there been any updates or current workarounds for this?

I was thinking of appending the UMI sequence to R1 and R2 using this method: OpenGene/fastp#23