As part of quality control, it's often desired to know how well reference-based OTU picking works for our samples by checking the percentage of reads passing the closed-reference OTU picking. Though it's not difficult to calculate it manually using tools available in QIIME2, it'd be great to have that information in the output of vsearch-cluster-features-closed-reference as well.
I agree, this could be useful even if it is easy to figure out with additional steps in QIIME2. I have raised an issue here and we will post back to this forum thread when we have updates.
I would like to second this feature suggestion I see on the github thread @Nicholas_Bokulich linked to that @colinbrislawn pointed out a way to do this in vsearch, but as far as I can tell that work-around involves exporting all the qiime2 artifacts and running the command in vsearch in order to get the log files which conveniently show the % of seqs passing.
I'm interested in this feature beacause I am working on a PICRUSt analysis from ASVs and would like to report what % of my sequences passed closed-ref clustering. I'm also trying to judge what -p-perc-identity to use for closed-ref.
that is correct ā @colinbrislawn's example is running vsearch directly.
Since we have that open issue that you found (and @colinbrislawn's very helpful advice for obtaining this info directly from vsearch), we should have that feature added to QIIME2 some day in the (hopefully) near future. Unfortunately, I do not have an eta though! Contributions are always very welcome .
Until then, you can check input (see demux summarize) and output (see feature-table summarize) sequence frequencies to determine the % that matched the reference.
Sorry we don't have an easier solution in place yet! Thank you for the feedback.