Number of features in ANCOM percentile abundance table is way less than in my feature table


Since the number of features (taxa) shown in my ANCOM percentile abundance table is only 193 and yet the corresponding feature table shows 1,145 features, what happened to the rest of the my features? Are the rest of the features identical in their percentile abundances across the two treatments and therefore are not being shown? Please advise. I have attached both .qzv files (from ANCOM and from the corresponding filtered feature table).

Thank you,

mesocosms-nonchimeric-w-borderline-Exper.Day14to20-table.qzv (617.7 KB)
Day14to20-Level7-ancom-Treatment_gg99_nonchimeric-w-b.qzv (726.5 KB)

Hi @Lisa_Crummett,
It appears that before running ANCOM you collapsed your feature table to the species “7” level. This collapsed your 1,145 features (ASVs) into 193 species which is what is being used and plotted for ANCOM.
If you want to stick at the ASV level, you should use the non-collapsed table.

1 Like

Thank you Mehrbod, that makes sense. So features that are unidentified at the species level but share the same genus are getting lumped together, right?


Hi @Lisa_Crummett
Correct! Though in some cases they may even get lumped to the family level or higher if the reference database used doesn’t have them differentiated at the genus level either.

1 Like

This topic was automatically closed 31 days after the last reply. New replies are no longer allowed.