Core features and taxonomy

[Alan_Chan (Alan)]

I am having an issue understanding some discrepancies using qiime feature-table core-features. When I run core features on a feature table that hasn't been combined with taxonomy yet, the number of core features is much different than if I were to run the script on a feature table that has been combined with taxonomy at the species level. When I try it with a feature table at the phylum level, it is also very different than the other two. Why are the number of features and core features changing depending on whether my feature table has been combined with taxonomy or not?

I'm attaching the qzv files in case that helps.
Here is the core feature without taxonomy core-otus.qzv (255.7 KB), with taxonomy at species level infectedcorefeatires.qzv (235.1 KB), and with taxonomy at phylum level core-phyla.qzv (235.7 KB).

[Nicholas_Bokulich]

Hi @Alan_Chan,

@Alan_Chan

When I run core features on a feature table that hasn’t been combined with taxonomy yet, the number of core features is much different than if I were to run the script on a feature table that has been combined with taxonomy at the species level.

Using qiime taxa collapse is not "combining" your feature table with the taxonomy, it is collapsing the feature table based on taxonomy... so any features (ASVs) that share the same taxonomic label at a given level are collapsed into a single feature. Hence, collapsing at phylum level will give very different results from collapsing at species level, and both will be very different from an uncollapsed feature table.