Good morning QIIME folks, The problem This question relates to looking at a barplot, realizing a taxonomic identifier is being labeled incorrectly, and trying to figure out why it's incorrectly labeled. The incorrect label is rare, however, and I can't find the label that is assigned in that barpl…

Hi @devonorourke , Interesting problem! Thank you for detailed info on the debugging you've done so far! This is almost certainly related to the custom reference database you've generated — those sequences must be in there somewhere. [image] devonorourke: I’ve tried looking back into the origi…

Hey @Nicholas_Bokulich , Thanks for the reply. Let me add two other observations after continuing to work on this problem: I ran the exact same dataset through the blast classifier, and the barplot doesn't have those two weird labels - they're gone, and no new weirdness has arisen. I ran the exac…

Hi @devonorourke , [image] devonorourke: I ran the exact same dataset through the same Vsearch classifier but added two additional parameters (altering the percent identity and number of max-accepts), and the weirdness was gone in the barplot again too! Very interesting... I did notice that …

Hi again @Nicholas_Bokulich , I didn't upload the .fa or .txt files to the repo because they're pretty big, but I did upload their .qza equivalents. You should be able to see everything here - this repo isn't private at the moment: guano/Mangan/data/qiimeoutput at master · devonorourke/guan…

Hi @devonorourke , The old and new vsearch barplots appear to have used different reference files — the reference artifacts in the repo are for the new barplot. What I can tell so far, though, is that those bizarre assignments have annotations down to species level, e.g.: k__GAATTAGGCCAACCAGGGGCCC…

Thanks @Nicholas_Bokulich , I'm not positive which artifact you are referring to as old and new to you - I'm assuming "old" was Mangan_bplot_vsearch.qzv , while "new" is " Mangan_bplot_vsearch_p90m50k.qzv ". The file with the p90m50k.qza extension is the "new" one to me, and the one which has none of …

[image] devonorourke: I’m assuming “old” was Mangan_bplot_vsearch.qzv , while “new” is " Mangan_bplot_vsearch_p90m50k.qzv ". Yes, I was using those same definitions. [image] devonorourke: …odd. and yeah, I’m moving on from this for now. I’ll let you know if it jumps up again. Thanks! Do…

I'm starting to think this would have been better with a phone call! Thanks @Nicholas_Bokulich for the detailed responses. Is this NCBI blast or classify-consensus-blast or standalone blastn? This phenomenon occurs with NCBI blast and classify-consensus blast. I haven't tested blastn from a lo…

[image] devonorourke: Is it possible for the documentation to be more explicit about the differences between these two classifiers yep, we have an open issue to clarify the difference. [image] devonorourke: I think we’re talking about two different parameters here. Maybe it was just a ty…

QIIME 2 Forum

Finding incorrectly formatted needle in apparently mostly formatted properly haystack

User Support