I’ve used the cutadapt plugin on previous workflows with success (such that I didn’t generate any error message), but recently noticed an error message posted by other users a few times. One such solution dated back in that halcyon era of March 2018 suggested to use standalone cutadapt and then presumably import that trimmed dataset into QIIME to continue with the workflow.
Comment and a question:
Comment:
I can confirm that by modifying the -m parameter in stand alone cutadapt that I was able to avoid generating any further error in my workflow.
Question:
Any chance that it would be possible to implement the addition of the -m parameter that @thermokarst suggested, switching the default minimum from 0 to 1, so that it avoids the error produced in QIIMEs cutadapt plugin described above?
Related: what’s the point in keeping a default value of 0?
You have my vote! The hold-up has been a bit of a technical one, that has to do with "empty" or missing data (and really this is a bigger problem than just some demuxed sequences, this is a bit of an existential question - "what do I represent when I have no content (soul)?"). My guess is that problem won't be solved for some time, so in the meantime, we should just make this happen in q2-cutadapt. I will move this ticket into the "planned" column of this release.
Because we generally stick with the defaults set by independent programs, since they were chosen by someone with more intimate knowledge to the problem than us (I know, I know, that is a bit of an assumption, but still...).
