Torbjørn Rognes

This also applies to the sortbysize and sortbylength commands. We could issue a warning, yes. An alternative is to extend the shuffle and sort commands to officially work with FASTQ...

Thanks for suggesting the idea of a banded alignment. This is something that we have and will consider. It is clear that full alignment is time-consuming with long sequences and...

related to #25

Or perhaps [parasail](https://github.com/jeffdaily/parasail)?

I like the idea of using Parasail, but after considering it a bit I am a bit concerned about the performance. Computing the alignments now accounts for about 80% of...

Thanks for reporting this issue. VSEARCH includes masked regions in the alignment score, as does USEARCH: http://drive5.com/usearch/manual/masking.html When searching or clustering, masking is only used to exclude masked words or...

I agree, but in this case VSEARCH will treat the masked regions similarly to USEARCH for compatibility reasons.

This issue is related to issue #487.

The option should be accepted. Will fix.

Fixed in commit b83e8b7797ad05ede4928e7e8216b6bbb1badf8a. Both the `minseqlength` and the `maxseqlength` option is allowed now.