Expose fastq_join command to concatenate (not merge) PE reads

[colinbrislawn]

Vsearch supports --fastq_join in which "sequences are not merged as with the fastq_mergepairs command, but simply joined with a gap."

Addition Description
It would work like the existing merge-pairs (q2 plugin docs).

Current Behavior
only merging is supported via overlap

Proposed Behavior
Add a new semantic type for formally-paired (now concatenated!) Illumina reads
Add warnings and docs about how this differs from most other methods

Questions

1. Who wants to make this new semantic type?
2. What programs take discontiguous amplicons as input?

Refs for vsearch
Forums x-ref.
803f-1392r amplicons is 590 reads, using PE300 x-ref

Refs for DADA2
https://forum.qiime2.org/t/dada2-option-justconcatenate/21661/2
https://forum.qiime2.org/t/classification-of-non-overlapping-reads-treated-with-justconcatenate/17914/7

[lizgehret]

Hey @colinbrislawn, thanks for submitting this! People have definitely been requesting this for a while, but this will be a big undertaking. We're going to discuss this in our engineering meeting next week to see what kind of timeline makes sense for getting this implemented!

[lizgehret]

Hey @colinbrislawn circling back on this - what are some of the actions you're hoping to utilize this on downstream? Just looking for a better scope for prioritizing this - this would be easy enough to create, but we won't be able to use it with anything as it currently stands.

[colinbrislawn]

we won't be able to use it with anything

Yeah, we also reached this conclusion last time. Discontiguous amplicons break a lot of assumptions, so we may want to put this on hold until we have the downstream pipeline planned.

(If someone has reads like this, they should comment here what they plan to do with them!)

[lizgehret]

Gotcha! Yeah agreed, we'll put this on hold for now and see if we receive any replies (or forum posts) in the coming months. I'm going to move this to our 'approved but unscheduled' project board in the meantime!