I’m following the Moving Pictures tutorial and go to the ANCOM analysis. When I looked at its visualization, I found that the two results (one significant genus in each subject) are actually the same genus! I checked the ID string in the taxonomy.qzv, searched for both 4b5eeb300368260019c1fbc7a3c718fc (present only in subject 1) and 868528ca947bc57b69ffdf83e6b73bae (present only in subject 2) and found that they are both k__Bacteria; p__Bacteroidetes; c__Bacteroidia; o__Bacteroidales; f__Bacteroidaceae; g__Bacteroides; s__. Is this a bug?
Hi @jjmmii! This isn’t unexpected: assuming you denoised your data with DADA2 or Deblur, each feature in your feature table represents a single Amplicon Sequence Variant (ASV), i.e. a biological sequence. Different ASVs can map to the same taxon (for example, the same genus) – in other words, ASVs don’t necessarily correspond to different taxa. The same idea applies to OTUs – for example, if you’ve clustered your sequences into 97% OTUs, different OTU representative sequences may map to the same genus.
The ANCOM section of the Moving Pictures tutorial shows how to perform differential abundance testing using ASVs (i.e. the results you linked to above), as well as with taxa at a specific taxonomic level. In the latter part of the tutorial, the table is collapsed to genus level and run through ANCOM, but you can collapse at whatever taxonomic level you’re interested in using qiime taxa collapse.
1 Like
Ah got it. Thank you so much!
1 Like
This topic was automatically closed 31 days after the last reply. New replies are no longer allowed.